bayes.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First we import the file and get it's contents"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import csv\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>participant_id</th>\n",
       "      <th>id</th>\n",
       "      <th>sex</th>\n",
       "      <th>recruit_date</th>\n",
       "      <th>postal_code</th>\n",
       "      <th>province</th>\n",
       "      <th>family_country</th>\n",
       "      <th>row_school</th>\n",
       "      <th>bus</th>\n",
       "      <th>sports_type</th>\n",
       "      <th>...</th>\n",
       "      <th>other</th>\n",
       "      <th>flu_binary</th>\n",
       "      <th>flu_date</th>\n",
       "      <th>vaccines_binary</th>\n",
       "      <th>comorbidities_complete</th>\n",
       "      <th>final_classification_of_th</th>\n",
       "      <th>survey_end_date</th>\n",
       "      <th>name_initials_of_the_inter</th>\n",
       "      <th>comments</th>\n",
       "      <th>final_outcome_complete</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>HVH-0001-78</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-11-09</td>\n",
       "      <td>08025</td>\n",
       "      <td>1.0</td>\n",
       "      <td>296.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-11-09</td>\n",
       "      <td>ASA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>HVH-0002-78</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-11-14</td>\n",
       "      <td>08012</td>\n",
       "      <td>1.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-11-14</td>\n",
       "      <td>ASA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>HVH-0003-78</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-11-16</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>314.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-11-16</td>\n",
       "      <td>ASA</td>\n",
       "      <td>Amigdalitis estreptocòcica</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>HVH-0004-78</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-11-30</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>416.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-12-01</td>\n",
       "      <td>ASA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>HVH-0005-78</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-12-16</td>\n",
       "      <td>08016</td>\n",
       "      <td>1.0</td>\n",
       "      <td>269.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-12-16</td>\n",
       "      <td>ASA</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1676</th>\n",
       "      <td>1220-65</td>\n",
       "      <td>90-0075-78</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-12-16</td>\n",
       "      <td>08210</td>\n",
       "      <td>1.0</td>\n",
       "      <td>421.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-12-18</td>\n",
       "      <td>DCV</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1677</th>\n",
       "      <td>1304-1</td>\n",
       "      <td>ESB-0001-78</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-12-14</td>\n",
       "      <td>08830</td>\n",
       "      <td>1.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ISABEL AGUILAR MOLINER</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1678</th>\n",
       "      <td>1304-2</td>\n",
       "      <td>ESB-0002-78</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-12-17</td>\n",
       "      <td>08830</td>\n",
       "      <td>1.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ISABEL AGUILAR MOLLINER</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1679</th>\n",
       "      <td>1304-3</td>\n",
       "      <td>ESB-0003-78</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2020-12-17</td>\n",
       "      <td>08830</td>\n",
       "      <td>1.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ISABEL AGUILAR MOLLINER</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1680</th>\n",
       "      <td>1304-4</td>\n",
       "      <td>ESB-0004-78</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2020-12-17</td>\n",
       "      <td>08830</td>\n",
       "      <td>1.0</td>\n",
       "      <td>302.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>12.0</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>ISABEL AGUILAR MOLLINER</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1681 rows × 197 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     participant_id           id  sex recruit_date postal_code  province  \\\n",
       "0                 1  HVH-0001-78  1.0   2020-11-09       08025       1.0   \n",
       "1                 2  HVH-0002-78  2.0   2020-11-14       08012       1.0   \n",
       "2                 3  HVH-0003-78  1.0   2020-11-16         NaN       1.0   \n",
       "3                 4  HVH-0004-78  2.0   2020-11-30         NaN       1.0   \n",
       "4                 5  HVH-0005-78  2.0   2020-12-16       08016       1.0   \n",
       "...             ...          ...  ...          ...         ...       ...   \n",
       "1676        1220-65   90-0075-78  1.0   2020-12-16       08210       1.0   \n",
       "1677         1304-1  ESB-0001-78  2.0   2020-12-14       08830       1.0   \n",
       "1678         1304-2  ESB-0002-78  1.0   2020-12-17       08830       1.0   \n",
       "1679         1304-3  ESB-0003-78  1.0   2020-12-17       08830       1.0   \n",
       "1680         1304-4  ESB-0004-78  2.0   2020-12-17       08830       1.0   \n",
       "\n",
       "      family_country row_school  bus  sports_type  ...  other flu_binary  \\\n",
       "0              296.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "1              302.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "2              314.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "3              416.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "4              269.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "...              ...        ...  ...          ...  ...    ...        ...   \n",
       "1676           421.0        NaN  NaN          NaN  ...    NaN        0.0   \n",
       "1677           302.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "1678           302.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "1679           302.0        NaN  2.0          NaN  ...    NaN        0.0   \n",
       "1680           302.0          2  2.0         12.0  ...    NaN        0.0   \n",
       "\n",
       "     flu_date vaccines_binary comorbidities_complete  \\\n",
       "0         NaN             1.0                      2   \n",
       "1         NaN             1.0                      2   \n",
       "2         NaN             1.0                      2   \n",
       "3         NaN             1.0                      2   \n",
       "4         NaN             1.0                      2   \n",
       "...       ...             ...                    ...   \n",
       "1676      NaN             1.0                      2   \n",
       "1677      NaN             1.0                      2   \n",
       "1678      NaN             1.0                      2   \n",
       "1679      NaN             1.0                      2   \n",
       "1680      NaN             1.0                      2   \n",
       "\n",
       "      final_classification_of_th  survey_end_date  name_initials_of_the_inter  \\\n",
       "0                            2.0       2020-11-09                         ASA   \n",
       "1                            2.0       2020-11-14                         ASA   \n",
       "2                            1.0       2020-11-16                         ASA   \n",
       "3                            2.0       2020-12-01                         ASA   \n",
       "4                            1.0       2020-12-16                         ASA   \n",
       "...                          ...              ...                         ...   \n",
       "1676                         1.0       2020-12-18                         DCV   \n",
       "1677                         NaN              NaN     ISABEL AGUILAR MOLINER    \n",
       "1678                         1.0              NaN    ISABEL AGUILAR MOLLINER    \n",
       "1679                         1.0              NaN    ISABEL AGUILAR MOLLINER    \n",
       "1680                         1.0              NaN    ISABEL AGUILAR MOLLINER    \n",
       "\n",
       "                        comments  final_outcome_complete  \n",
       "0                            NaN                       2  \n",
       "1                            NaN                       2  \n",
       "2     Amigdalitis estreptocòcica                       2  \n",
       "3                            NaN                       2  \n",
       "4                            NaN                       2  \n",
       "...                          ...                     ...  \n",
       "1676                         NaN                       2  \n",
       "1677                         NaN                       0  \n",
       "1678                         NaN                       0  \n",
       "1679                         NaN                       0  \n",
       "1680                         NaN                       0  \n",
       "\n",
       "[1681 rows x 197 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv(\"./data.csv\")\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "scrolled": false
   },
   "source": [
    "Then we make a first clean of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('participant_id', axis=1)\n",
    "df = df.drop('recruit_date', axis=1)\n",
    "df = df.drop('postal_code', axis=1)\n",
    "df = df.drop('province', axis=1)\n",
    "df = df.drop('family_country', axis=1)\n",
    "df = df.drop('row_school', axis=1)\n",
    "df = df.drop('floor_level', axis=1)\n",
    "df = df.drop('survey_type', axis=1)\n",
    "df = df.drop('inclusion_criteria', axis=1)\n",
    "df = df.drop('housemember_symptoms___1', axis=1)\n",
    "df = df.drop('housemember_symptoms___2', axis=1)\n",
    "df = df.drop('housemember_symptoms___3', axis=1)\n",
    "df = df.drop('housemember_symptoms___4', axis=1)\n",
    "df = df.drop('housemember_symptoms___5', axis=1)\n",
    "df = df.drop('persons_home', axis=1)\n",
    "df = df.drop('simptomatology_date', axis=1)\n",
    "df = df.drop('date_fever', axis=1)\n",
    "df = df.drop('end_fever', axis=1)\n",
    "df = df.drop('other_simptomatology_lliure', axis=1)\n",
    "df = df.drop('sero_type_response', axis=1)\n",
    "df = df.drop('sero_method', axis=1)\n",
    "df = df.drop('sero_date', axis=1)\n",
    "df = df.drop('sero', axis=1)\n",
    "df = df.drop('cxr', axis=1)\n",
    "df = df.drop('ct', axis=1)\n",
    "df = df.drop('cxr_date', axis=1)\n",
    "df = df.drop('thoracic_ct_date', axis=1)\n",
    "df = df.drop('pcr_performed', axis=1)\n",
    "df = df.drop('pcr_date', axis=1)\n",
    "df = df.drop('pcr_type', axis=1)\n",
    "df = df.drop('antigenic_performed', axis=1)\n",
    "df = df.drop('ag_test_mark', axis=1)\n",
    "df = df.drop('antigenic_date', axis=1)\n",
    "df = df.drop('antigenic_sample', axis=1)\n",
    "df = df.drop('othervirus_performed', axis=1)\n",
    "df = df.drop('otherviruses_date', axis=1)\n",
    "df = df.drop('vrs_result', axis=1)\n",
    "df = df.drop('adeno_result', axis=1)\n",
    "df = df.drop('flu_a_result', axis=1)\n",
    "df = df.drop('flu_b_result', axis=1)\n",
    "df = df.drop('adm_date', axis=1)\n",
    "df = df.drop('picu_adm', axis=1)\n",
    "df = df.drop('adm_hospital', axis=1)\n",
    "df = df.drop('discharge_date', axis=1)\n",
    "df = df.drop('final_diagnosis_text', axis=1)\n",
    "df = df.drop('coviral_binary', axis=1)\n",
    "df = df.drop('coviral_type', axis=1)\n",
    "df = df.drop('other_viruses_text', axis=1)\n",
    "df = df.drop('bacterial_infection', axis=1)\n",
    "df = df.drop('bacterial_type', axis=1)\n",
    "df = df.drop('clinical_and_diagnosis_data_at_the_admission_time_complete', axis=1)\n",
    "list_of_shit = ['comorbi_binary',\n",
    " 'cardiopathy___1',\n",
    " 'cardiopathy___2',\n",
    " 'cardiopathy___3',\n",
    " 'hypertension___1',\n",
    " 'hypertension___2',\n",
    " 'hypertension___3',\n",
    " 'pulmonar_disease___1',\n",
    " 'pulmonar_disease___2',\n",
    " 'pulmonar_disease___3',\n",
    " 'asma___1',\n",
    " 'asma___2',\n",
    " 'asma___3',\n",
    " 'nephrology___1',\n",
    " 'nephrology___2',\n",
    " 'nephrology___3',\n",
    " 'hepatic___1',\n",
    " 'hepatic___2',\n",
    " 'hepatic___3',\n",
    " 'neurologic___1',\n",
    " 'neurologic___2',\n",
    " 'neurologic___3',\n",
    " 'diabetes___1',\n",
    " 'diabetes___2',\n",
    " 'diabetes___3',\n",
    " 'tuberculosi___1',\n",
    " 'tuberculosi___2',\n",
    " 'tuberculosi___3',\n",
    " 'idp___1',\n",
    " 'idp___2',\n",
    " 'idp___3',\n",
    " 'neoplasia___1',\n",
    " 'neoplasia___2',\n",
    " 'neoplasia___3',\n",
    " 'kawasaki___1',\n",
    " 'kawasaki___2',\n",
    " 'kawasaki___3',\n",
    " 'inflammation___1',\n",
    " 'inflammation___2',\n",
    " 'inflammation___3',\n",
    " 'vih_others___1',\n",
    " 'vih_others___2',\n",
    " 'vih_others___3',\n",
    " 'obesity',\n",
    " 'other']\n",
    "\n",
    "for shit in list_of_shit:\n",
    "    df = df.drop(shit, axis=1)\n",
    "\n",
    "df = df.drop('flu_date', axis=1)\n",
    "df = df.drop('vaccines_binary', axis=1)\n",
    "df = df.drop('comorbidities_complete', axis=1)\n",
    "df = df.drop('survey_end_date', axis=1)\n",
    "df = df.drop('name_initials_of_the_inter', axis=1)\n",
    "df = df.drop('comments', axis=1)\n",
    "df = df.drop('final_outcome_complete', axis=1)\n",
    "df = df.drop('id', axis=1)\n",
    "df = df.drop('bus', axis=1)\n",
    "df = df.drop('sports_type', axis=1)\n",
    "df = df.drop('sports', axis=1)\n",
    "df = df.drop('m2', axis=1)\n",
    "df = df.drop('rooms', axis=1)\n",
    "df = df.drop('school_symptoms_member___1', axis=1)\n",
    "df = df.drop('school_symptoms_member___2', axis=1)\n",
    "df = df.drop('school_symptoms_member___4', axis=1)\n",
    "df = df.drop('school_symptoms_member___5', axis=1)\n",
    "df = df.drop('sero_type_response_2', axis=1)\n",
    "df = df.drop('sat_hb_o2', axis=1)\n",
    "df = df.drop('final_outcome', axis=1)\n",
    "df = df.drop('final_classification_of_th', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['sex',\n",
       " 'smokers_home',\n",
       " 'sympt_epi',\n",
       " 'home_confirmed',\n",
       " 'school_symptoms',\n",
       " 'school_confirmed',\n",
       " 'symptoms_binary',\n",
       " 'fever',\n",
       " 'highest_fever',\n",
       " 'total_days_fever',\n",
       " 'tos',\n",
       " 'cough_first',\n",
       " 'crup',\n",
       " 'crup_first',\n",
       " 'dysphonia',\n",
       " 'disfonia_first',\n",
       " 'resp',\n",
       " 'dyspnea_first',\n",
       " 'tachypnea',\n",
       " 'tachypnea_first',\n",
       " 'ausc_resp',\n",
       " 'auscult_first',\n",
       " 'wheezing',\n",
       " 'crackles',\n",
       " 'odynophagia',\n",
       " 'odynophagia_first',\n",
       " 'nasal_congestion',\n",
       " 'nasal_first',\n",
       " 'fatiga',\n",
       " 'fatigue_first',\n",
       " 'headache',\n",
       " 'headache_first',\n",
       " 'conjuntivitis',\n",
       " 'conj_first',\n",
       " 'ocular_pain',\n",
       " 'ocular_first',\n",
       " 'gi_symptoms',\n",
       " 'gi_first',\n",
       " 'abdominal_pain',\n",
       " 'vomiting',\n",
       " 'dyarrea',\n",
       " 'dermatologic',\n",
       " 'skin_first',\n",
       " 'rash',\n",
       " 'inflam_periferic',\n",
       " 'inflam_oral',\n",
       " 'adenopathies',\n",
       " 'lymph_first',\n",
       " 'hepato',\n",
       " 'hepato_first',\n",
       " 'splenomegaly',\n",
       " 'spleno_first',\n",
       " 'hemorrhagies',\n",
       " 'hemorr_first',\n",
       " 'irritability',\n",
       " 'irritability_first',\n",
       " 'neuro',\n",
       " 'neuro_first',\n",
       " 'confusion',\n",
       " 'seizures',\n",
       " 'nuchal_stiffness',\n",
       " 'hypotonia',\n",
       " 'peripheral_paralysis',\n",
       " 'shock',\n",
       " 'shock_first',\n",
       " 'taste_smell',\n",
       " 'taste_first',\n",
       " 'smell',\n",
       " 'smell_first',\n",
       " 'other_simptomatology',\n",
       " 'other_simptomatology_3',\n",
       " 'other_simptomatology_4',\n",
       " 'sero_response',\n",
       " 'cxr2',\n",
       " 'ct2',\n",
       " 'sat_hb_o2_value',\n",
       " 'pcr_result',\n",
       " 'antigenic_result',\n",
       " 'final_diagnosis_code',\n",
       " 'flu_binary']"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sex</th>\n",
       "      <th>smokers_home</th>\n",
       "      <th>sympt_epi</th>\n",
       "      <th>home_confirmed</th>\n",
       "      <th>school_symptoms</th>\n",
       "      <th>school_confirmed</th>\n",
       "      <th>symptoms_binary</th>\n",
       "      <th>fever</th>\n",
       "      <th>highest_fever</th>\n",
       "      <th>total_days_fever</th>\n",
       "      <th>...</th>\n",
       "      <th>other_simptomatology_3</th>\n",
       "      <th>other_simptomatology_4</th>\n",
       "      <th>sero_response</th>\n",
       "      <th>cxr2</th>\n",
       "      <th>ct2</th>\n",
       "      <th>sat_hb_o2_value</th>\n",
       "      <th>pcr_result</th>\n",
       "      <th>antigenic_result</th>\n",
       "      <th>final_diagnosis_code</th>\n",
       "      <th>flu_binary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>92</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>98</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1676</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1677</th>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1678</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1679</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1680</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1681 rows × 80 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      sex  smokers_home  sympt_epi  home_confirmed  school_symptoms  \\\n",
       "0     1.0           2.0        0.0             0.0              0.0   \n",
       "1     2.0           2.0        0.0             0.0              0.0   \n",
       "2     1.0           2.0        0.0             0.0              0.0   \n",
       "3     2.0           2.0        0.0             0.0              NaN   \n",
       "4     2.0           2.0        1.0             1.0              NaN   \n",
       "...   ...           ...        ...             ...              ...   \n",
       "1676  1.0           NaN        0.0             0.0              0.0   \n",
       "1677  2.0           NaN        0.0             0.0              0.0   \n",
       "1678  1.0           2.0        0.0             0.0              0.0   \n",
       "1679  1.0           2.0        0.0             0.0              0.0   \n",
       "1680  2.0           1.0        0.0             0.0              1.0   \n",
       "\n",
       "      school_confirmed  symptoms_binary  fever  highest_fever  \\\n",
       "0                  0.0              1.0    1.0            1.0   \n",
       "1                  0.0              1.0    1.0            2.0   \n",
       "2                  0.0              1.0    1.0            2.0   \n",
       "3                  NaN              1.0    2.0            NaN   \n",
       "4                  NaN              1.0    1.0            3.0   \n",
       "...                ...              ...    ...            ...   \n",
       "1676               0.0              1.0    1.0            2.0   \n",
       "1677               0.0              1.0    1.0            1.0   \n",
       "1678               0.0              1.0    1.0            2.0   \n",
       "1679               0.0              1.0    2.0            NaN   \n",
       "1680               1.0              1.0    1.0            1.0   \n",
       "\n",
       "     total_days_fever  ...  other_simptomatology_3  other_simptomatology_4  \\\n",
       "0                   1  ...                     NaN                     NaN   \n",
       "1                   1  ...                     NaN                     NaN   \n",
       "2                   3  ...                     NaN                     NaN   \n",
       "3                 NaN  ...                     NaN                     NaN   \n",
       "4                   1  ...                     NaN                     NaN   \n",
       "...               ...  ...                     ...                     ...   \n",
       "1676                2  ...                     NaN                     NaN   \n",
       "1677              NaN  ...                     NaN                     NaN   \n",
       "1678              NaN  ...                     NaN                     NaN   \n",
       "1679              NaN  ...                     NaN                     NaN   \n",
       "1680                1  ...                     NaN                     NaN   \n",
       "\n",
       "      sero_response  cxr2  ct2  sat_hb_o2_value  pcr_result  antigenic_result  \\\n",
       "0               NaN   NaN  NaN               92         NaN               2.0   \n",
       "1               NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "2               NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "3               NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "4               NaN   NaN  NaN               98         2.0               NaN   \n",
       "...             ...   ...  ...              ...         ...               ...   \n",
       "1676            NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "1677            NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "1678            NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "1679            NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "1680            NaN   NaN  NaN              NaN         NaN               2.0   \n",
       "\n",
       "      final_diagnosis_code  flu_binary  \n",
       "0                      3.0         0.0  \n",
       "1                      3.0         0.0  \n",
       "2                      3.0         0.0  \n",
       "3                      3.0         0.0  \n",
       "4                      3.0         0.0  \n",
       "...                    ...         ...  \n",
       "1676                   3.0         0.0  \n",
       "1677                   3.0         0.0  \n",
       "1678                   3.0         0.0  \n",
       "1679                   3.0         0.0  \n",
       "1680                   3.0         0.0  \n",
       "\n",
       "[1681 rows x 80 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We check the number of Nan values for each column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "not_null = lambda x : x.notnull()\n",
    "def list_nan_values(df, function):\n",
    "    check_nan = lambda x : len(df) - len(df[function(x)])\n",
    "    list_nan_cols = list()\n",
    "    for col in df.columns:\n",
    "        list_nan_cols.append((col, check_nan(df[col])))\n",
    "    return list_nan_cols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "list_nan = list_nan_values(df, not_null)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('sex', 4),\n",
       " ('smokers_home', 156),\n",
       " ('sympt_epi', 17),\n",
       " ('home_confirmed', 151),\n",
       " ('school_symptoms', 51),\n",
       " ('school_confirmed', 182),\n",
       " ('symptoms_binary', 9),\n",
       " ('fever', 36),\n",
       " ('highest_fever', 699),\n",
       " ('total_days_fever', 827),\n",
       " ('tos', 38),\n",
       " ('cough_first', 881),\n",
       " ('crup', 890),\n",
       " ('crup_first', 1657),\n",
       " ('dysphonia', 44),\n",
       " ('disfonia_first', 1602),\n",
       " ('resp', 43),\n",
       " ('dyspnea_first', 1549),\n",
       " ('tachypnea', 43),\n",
       " ('tachypnea_first', 1604),\n",
       " ('ausc_resp', 328),\n",
       " ('auscult_first', 1526),\n",
       " ('wheezing', 1523),\n",
       " ('crackles', 1528),\n",
       " ('odynophagia', 49),\n",
       " ('odynophagia_first', 1278),\n",
       " ('nasal_congestion', 49),\n",
       " ('nasal_first', 837),\n",
       " ('fatiga', 47),\n",
       " ('fatigue_first', 1322),\n",
       " ('headache', 51),\n",
       " ('headache_first', 1255),\n",
       " ('conjuntivitis', 48),\n",
       " ('conj_first', 1648),\n",
       " ('ocular_pain', 778),\n",
       " ('ocular_first', 1675),\n",
       " ('gi_symptoms', 41),\n",
       " ('gi_first', 1146),\n",
       " ('abdominal_pain', 1147),\n",
       " ('vomiting', 1145),\n",
       " ('dyarrea', 1138),\n",
       " ('dermatologic', 48),\n",
       " ('skin_first', 1650),\n",
       " ('rash', 1650),\n",
       " ('inflam_periferic', 1649),\n",
       " ('inflam_oral', 1649),\n",
       " ('adenopathies', 43),\n",
       " ('lymph_first', 1652),\n",
       " ('hepato', 42),\n",
       " ('hepato_first', 1681),\n",
       " ('splenomegaly', 49),\n",
       " ('spleno_first', 1681),\n",
       " ('hemorrhagies', 51),\n",
       " ('hemorr_first', 1678),\n",
       " ('irritability', 1002),\n",
       " ('irritability_first', 1644),\n",
       " ('neuro', 43),\n",
       " ('neuro_first', 1671),\n",
       " ('confusion', 1671),\n",
       " ('seizures', 1671),\n",
       " ('nuchal_stiffness', 1671),\n",
       " ('hypotonia', 1671),\n",
       " ('peripheral_paralysis', 1671),\n",
       " ('shock', 119),\n",
       " ('shock_first', 1680),\n",
       " ('taste_smell', 219),\n",
       " ('taste_first', 1652),\n",
       " ('smell', 223),\n",
       " ('smell_first', 1652),\n",
       " ('other_simptomatology', 1616),\n",
       " ('other_simptomatology_3', 1664),\n",
       " ('other_simptomatology_4', 1679),\n",
       " ('sero_response', 1672),\n",
       " ('cxr2', 1633),\n",
       " ('ct2', 1666),\n",
       " ('sat_hb_o2_value', 1405),\n",
       " ('pcr_result', 1405),\n",
       " ('antigenic_result', 186),\n",
       " ('final_diagnosis_code', 66),\n",
       " ('flu_binary', 16)]"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list_nan"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We check wich vaues contain every column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sex': {1.0: 934, 2.0: 743, nan: 4},\n",
       " 'smokers_home': {2.0: 1074, 1.0: 451, nan: 156},\n",
       " 'sympt_epi': {0.0: 1370, 1.0: 294, nan: 17},\n",
       " 'home_confirmed': {0.0: 1411, 1.0: 119, nan: 151},\n",
       " 'school_symptoms': {0.0: 1531, 1.0: 99, nan: 51},\n",
       " 'school_confirmed': {0.0: 1415, 1.0: 84, nan: 182},\n",
       " 'symptoms_binary': {1.0: 1647, 0.0: 25, nan: 9},\n",
       " 'fever': {1.0: 983, 2.0: 662, nan: 36},\n",
       " 'highest_fever': {1.0: 285, 2.0: 535, 3.0: 162, nan: 699},\n",
       " 'total_days_fever': {'1': 285,\n",
       "  '3': 129,\n",
       "  '4': 82,\n",
       "  '2': 279,\n",
       "  '14': 1,\n",
       "  '1 DETERMINACION': 1,\n",
       "  '5': 22,\n",
       "  '2-3': 1,\n",
       "  '48': 1,\n",
       "  '6': 11,\n",
       "  '1 y medio': 1,\n",
       "  '10': 3,\n",
       "  '1,5': 3,\n",
       "  '38.7': 1,\n",
       "  '7': 5,\n",
       "  '24 hores': 3,\n",
       "  '24hores': 1,\n",
       "  '24h': 2,\n",
       "  '24horews': 1,\n",
       "  'va iniciar febreta a dies alterns una setmana abans, inferior a 37,4. El dia abans de consultar 37,6 mantingut': 1,\n",
       "  '16': 1,\n",
       "  '0,5': 4,\n",
       "  '2 dies ': 2,\n",
       "  '3 dies': 1,\n",
       "  '1 dia': 1,\n",
       "  '1 dia ': 2,\n",
       "  '2 DIES ': 1,\n",
       "  '4 dies': 1,\n",
       "  '0': 1,\n",
       "  '8': 1,\n",
       "  '9': 1,\n",
       "  '2,5': 4,\n",
       "  '7dias': 1,\n",
       "  nan: 827},\n",
       " 'tos': {1.0: 803, 2.0: 840, nan: 38},\n",
       " 'cough_first': {1.0: 736, 0.0: 64, nan: 881},\n",
       " 'crup': {2.0: 766, 1.0: 24, 3.0: 1, nan: 890},\n",
       " 'crup_first': {0.0: 7, 1.0: 17, nan: 1657},\n",
       " 'dysphonia': {2.0: 1544, 1.0: 79, 3.0: 14, nan: 44},\n",
       " 'disfonia_first': {0.0: 10, 1.0: 69, nan: 1602},\n",
       " 'resp': {1.0: 133, 2.0: 1504, 3.0: 1, nan: 43},\n",
       " 'dyspnea_first': {1.0: 112, 0.0: 20, nan: 1549},\n",
       " 'tachypnea': {1.0: 78, 2.0: 1550, 3.0: 10, nan: 43},\n",
       " 'tachypnea_first': {1.0: 66, 0.0: 11, nan: 1604},\n",
       " 'ausc_resp': {2.0: 158, 1.0: 1195, nan: 328},\n",
       " 'auscult_first': {1.0: 126, 0.0: 29, nan: 1526},\n",
       " 'wheezing': {1.0: 140, 2.0: 18, nan: 1523},\n",
       " 'crackles': {2.0: 107, 1.0: 46, nan: 1528},\n",
       " 'odynophagia': {2.0: 1120, 1.0: 406, 3.0: 106, nan: 49},\n",
       " 'odynophagia_first': {1.0: 385, 0.0: 18, nan: 1278},\n",
       " 'nasal_congestion': {1.0: 849, 2.0: 779, 3.0: 4, nan: 49},\n",
       " 'nasal_first': {1.0: 794, 0.0: 50, nan: 837},\n",
       " 'fatiga': {2.0: 1225, 1.0: 362, 3.0: 47, nan: 47},\n",
       " 'fatigue_first': {0.0: 21, 1.0: 338, nan: 1322},\n",
       " 'headache': {2.0: 1045, 3.0: 156, 1.0: 429, nan: 51},\n",
       " 'headache_first': {1.0: 401, 0.0: 25, nan: 1255},\n",
       " 'conjuntivitis': {2.0: 1594, 1.0: 33, 3.0: 6, nan: 48},\n",
       " 'conj_first': {1.0: 26, 0.0: 7, nan: 1648},\n",
       " 'ocular_pain': {3.0: 146, 2.0: 751, 1.0: 6, nan: 778},\n",
       " 'ocular_first': {1.0: 4, 0.0: 2, nan: 1675},\n",
       " 'gi_symptoms': {0.0: 1096, 1.0: 544, nan: 41},\n",
       " 'gi_first': {1.0: 493, 0.0: 42, nan: 1146},\n",
       " 'abdominal_pain': {2.0: 203, 1.0: 307, 3.0: 24, nan: 1147},\n",
       " 'vomiting': {1.0: 236, 2.0: 300, nan: 1145},\n",
       " 'dyarrea': {2.0: 234, 1.0: 308, 3.0: 1, nan: 1138},\n",
       " 'dermatologic': {0.0: 1601, 1.0: 32, nan: 48},\n",
       " 'skin_first': {0.0: 14, 1.0: 17, nan: 1650},\n",
       " 'rash': {1.0: 25, 0.0: 6, nan: 1650},\n",
       " 'inflam_periferic': {0.0: 31, 1.0: 1, nan: 1649},\n",
       " 'inflam_oral': {0.0: 31, 1.0: 1, nan: 1649},\n",
       " 'adenopathies': {2.0: 1267, 1.0: 32, 3.0: 339, nan: 43},\n",
       " 'lymph_first': {1.0: 27, 0.0: 2, nan: 1652},\n",
       " 'hepato': {2.0: 1294, 3.0: 345, nan: 42},\n",
       " 'hepato_first': {nan: 1681},\n",
       " 'splenomegaly': {2.0: 1289, 3.0: 343, nan: 49},\n",
       " 'spleno_first': {nan: 1681},\n",
       " 'hemorrhagies': {2.0: 1496, 3.0: 131, 1.0: 3, nan: 51},\n",
       " 'hemorr_first': {1.0: 3, nan: 1678},\n",
       " 'irritability': {1.0: 37, 2.0: 632, 3.0: 10, nan: 1002},\n",
       " 'irritability_first': {1.0: 30, 0.0: 7, nan: 1644},\n",
       " 'neuro': {0.0: 1628, 1.0: 10, nan: 43},\n",
       " 'neuro_first': {1.0: 8, 0.0: 2, nan: 1671},\n",
       " 'confusion': {1.0: 1, 0.0: 9, nan: 1671},\n",
       " 'seizures': {0.0: 2, 1.0: 8, nan: 1671},\n",
       " 'nuchal_stiffness': {0.0: 10, nan: 1671},\n",
       " 'hypotonia': {1.0: 2, 0.0: 8, nan: 1671},\n",
       " 'peripheral_paralysis': {0.0: 10, nan: 1671},\n",
       " 'shock': {0.0: 1561, 1.0: 1, nan: 119},\n",
       " 'shock_first': {1.0: 1, nan: 1680},\n",
       " 'taste_smell': {0.0: 1432, 1.0: 30, nan: 219},\n",
       " 'taste_first': {0.0: 12, 1.0: 17, nan: 1652},\n",
       " 'smell': {0.0: 1427, 1.0: 31, nan: 223},\n",
       " 'smell_first': {0.0: 12, 1.0: 17, nan: 1652},\n",
       " 'other_simptomatology': {4.0: 33, 1.0: 17, 2.0: 15, nan: 1616},\n",
       " 'other_simptomatology_3': {4.0: 11, 2.0: 6, nan: 1664},\n",
       " 'other_simptomatology_4': {1.0: 1, 4.0: 1, nan: 1679},\n",
       " 'sero_response': {2.0: 5, 1.0: 4, nan: 1672},\n",
       " 'cxr2': {2.0: 35, 1.0: 13, nan: 1633},\n",
       " 'ct2': {2.0: 15, nan: 1666},\n",
       " 'sat_hb_o2_value': {'92': 3,\n",
       "  '98': 54,\n",
       "  '96': 7,\n",
       "  '99': 35,\n",
       "  '97': 35,\n",
       "  '100': 38,\n",
       "  '89': 4,\n",
       "  '97%': 13,\n",
       "  '96%': 3,\n",
       "  '98%': 28,\n",
       "  '95': 11,\n",
       "  '96-97': 1,\n",
       "  '100%': 10,\n",
       "  '99%': 11,\n",
       "  '90': 4,\n",
       "  '94': 5,\n",
       "  '91-92': 1,\n",
       "  '93': 2,\n",
       "  '91': 3,\n",
       "  '89-90%': 2,\n",
       "  '92%': 1,\n",
       "  '89%': 1,\n",
       "  '96% EL DIA 5/11': 1,\n",
       "  '99 %': 1,\n",
       "  '95-96': 1,\n",
       "  '90%': 1,\n",
       "  nan: 1405},\n",
       " 'pcr_result': {2.0: 192, 1.0: 84, nan: 1405},\n",
       " 'antigenic_result': {2.0: 1393, 1.0: 102, nan: 186},\n",
       " 'final_diagnosis_code': {3.0: 1362, 1.0: 178, 2.0: 75, nan: 66},\n",
       " 'flu_binary': {0.0: 1566, 1.0: 99, nan: 16}}"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import math\n",
    "def get_population(df, column):\n",
    "    rs = {}\n",
    "    values = list(column.dropna().unique())\n",
    "    for val in values:\n",
    "        rs[val] = len(df[column == val])\n",
    "    rs[np.nan] = len(column.isnull()[lambda x: x == True])\n",
    "    return rs\n",
    "check_values = {}\n",
    "for col in df.columns:\n",
    "    check_values[col] = get_population(df,df[col])\n",
    "check_values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We clean some other values from the dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('total_days_fever', axis=1)\n",
    "df = df.drop('sat_hb_o2_value', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sex': {1.0: 934, 2.0: 743, nan: 4},\n",
       " 'smokers_home': {2.0: 1074, 1.0: 451, nan: 156},\n",
       " 'sympt_epi': {0.0: 1370, 1.0: 294, nan: 17},\n",
       " 'home_confirmed': {0.0: 1411, 1.0: 119, nan: 151},\n",
       " 'school_symptoms': {0.0: 1531, 1.0: 99, nan: 51},\n",
       " 'school_confirmed': {0.0: 1415, 1.0: 84, nan: 182},\n",
       " 'symptoms_binary': {1.0: 1647, 0.0: 25, nan: 9},\n",
       " 'fever': {1.0: 983, 2.0: 662, nan: 36},\n",
       " 'highest_fever': {1.0: 285, 2.0: 535, 3.0: 162, nan: 699},\n",
       " 'tos': {1.0: 803, 2.0: 840, nan: 38},\n",
       " 'cough_first': {1.0: 736, 0.0: 64, nan: 881},\n",
       " 'crup': {2.0: 766, 1.0: 24, 3.0: 1, nan: 890},\n",
       " 'crup_first': {0.0: 7, 1.0: 17, nan: 1657},\n",
       " 'dysphonia': {2.0: 1544, 1.0: 79, 3.0: 14, nan: 44},\n",
       " 'disfonia_first': {0.0: 10, 1.0: 69, nan: 1602},\n",
       " 'resp': {1.0: 133, 2.0: 1504, 3.0: 1, nan: 43},\n",
       " 'dyspnea_first': {1.0: 112, 0.0: 20, nan: 1549},\n",
       " 'tachypnea': {1.0: 78, 2.0: 1550, 3.0: 10, nan: 43},\n",
       " 'tachypnea_first': {1.0: 66, 0.0: 11, nan: 1604},\n",
       " 'ausc_resp': {2.0: 158, 1.0: 1195, nan: 328},\n",
       " 'auscult_first': {1.0: 126, 0.0: 29, nan: 1526},\n",
       " 'wheezing': {1.0: 140, 2.0: 18, nan: 1523},\n",
       " 'crackles': {2.0: 107, 1.0: 46, nan: 1528},\n",
       " 'odynophagia': {2.0: 1120, 1.0: 406, 3.0: 106, nan: 49},\n",
       " 'odynophagia_first': {1.0: 385, 0.0: 18, nan: 1278},\n",
       " 'nasal_congestion': {1.0: 849, 2.0: 779, 3.0: 4, nan: 49},\n",
       " 'nasal_first': {1.0: 794, 0.0: 50, nan: 837},\n",
       " 'fatiga': {2.0: 1225, 1.0: 362, 3.0: 47, nan: 47},\n",
       " 'fatigue_first': {0.0: 21, 1.0: 338, nan: 1322},\n",
       " 'headache': {2.0: 1045, 3.0: 156, 1.0: 429, nan: 51},\n",
       " 'headache_first': {1.0: 401, 0.0: 25, nan: 1255},\n",
       " 'conjuntivitis': {2.0: 1594, 1.0: 33, 3.0: 6, nan: 48},\n",
       " 'conj_first': {1.0: 26, 0.0: 7, nan: 1648},\n",
       " 'ocular_pain': {3.0: 146, 2.0: 751, 1.0: 6, nan: 778},\n",
       " 'ocular_first': {1.0: 4, 0.0: 2, nan: 1675},\n",
       " 'gi_symptoms': {0.0: 1096, 1.0: 544, nan: 41},\n",
       " 'gi_first': {1.0: 493, 0.0: 42, nan: 1146},\n",
       " 'abdominal_pain': {2.0: 203, 1.0: 307, 3.0: 24, nan: 1147},\n",
       " 'vomiting': {1.0: 236, 2.0: 300, nan: 1145},\n",
       " 'dyarrea': {2.0: 234, 1.0: 308, 3.0: 1, nan: 1138},\n",
       " 'dermatologic': {0.0: 1601, 1.0: 32, nan: 48},\n",
       " 'skin_first': {0.0: 14, 1.0: 17, nan: 1650},\n",
       " 'rash': {1.0: 25, 0.0: 6, nan: 1650},\n",
       " 'inflam_periferic': {0.0: 31, 1.0: 1, nan: 1649},\n",
       " 'inflam_oral': {0.0: 31, 1.0: 1, nan: 1649},\n",
       " 'adenopathies': {2.0: 1267, 1.0: 32, 3.0: 339, nan: 43},\n",
       " 'lymph_first': {1.0: 27, 0.0: 2, nan: 1652},\n",
       " 'hepato': {2.0: 1294, 3.0: 345, nan: 42},\n",
       " 'hepato_first': {nan: 1681},\n",
       " 'splenomegaly': {2.0: 1289, 3.0: 343, nan: 49},\n",
       " 'spleno_first': {nan: 1681},\n",
       " 'hemorrhagies': {2.0: 1496, 3.0: 131, 1.0: 3, nan: 51},\n",
       " 'hemorr_first': {1.0: 3, nan: 1678},\n",
       " 'irritability': {1.0: 37, 2.0: 632, 3.0: 10, nan: 1002},\n",
       " 'irritability_first': {1.0: 30, 0.0: 7, nan: 1644},\n",
       " 'neuro': {0.0: 1628, 1.0: 10, nan: 43},\n",
       " 'neuro_first': {1.0: 8, 0.0: 2, nan: 1671},\n",
       " 'confusion': {1.0: 1, 0.0: 9, nan: 1671},\n",
       " 'seizures': {0.0: 2, 1.0: 8, nan: 1671},\n",
       " 'nuchal_stiffness': {0.0: 10, nan: 1671},\n",
       " 'hypotonia': {1.0: 2, 0.0: 8, nan: 1671},\n",
       " 'peripheral_paralysis': {0.0: 10, nan: 1671},\n",
       " 'shock': {0.0: 1561, 1.0: 1, nan: 119},\n",
       " 'shock_first': {1.0: 1, nan: 1680},\n",
       " 'taste_smell': {0.0: 1432, 1.0: 30, nan: 219},\n",
       " 'taste_first': {0.0: 12, 1.0: 17, nan: 1652},\n",
       " 'smell': {0.0: 1427, 1.0: 31, nan: 223},\n",
       " 'smell_first': {0.0: 12, 1.0: 17, nan: 1652},\n",
       " 'other_simptomatology': {4.0: 33, 1.0: 17, 2.0: 15, nan: 1616},\n",
       " 'other_simptomatology_3': {4.0: 11, 2.0: 6, nan: 1664},\n",
       " 'other_simptomatology_4': {1.0: 1, 4.0: 1, nan: 1679},\n",
       " 'sero_response': {2.0: 5, 1.0: 4, nan: 1672},\n",
       " 'cxr2': {2.0: 35, 1.0: 13, nan: 1633},\n",
       " 'ct2': {2.0: 15, nan: 1666},\n",
       " 'pcr_result': {2.0: 192, 1.0: 84, nan: 1405},\n",
       " 'antigenic_result': {2.0: 1393, 1.0: 102, nan: 186},\n",
       " 'final_diagnosis_code': {3.0: 1362, 1.0: 178, 2.0: 75, nan: 66},\n",
       " 'flu_binary': {0.0: 1566, 1.0: 99, nan: 16}}"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "check_values = {}\n",
    "for col in df.columns:\n",
    "    check_values[col] = get_population(df,df[col])\n",
    "check_values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We do a heatmap in order to see the correlation between variables, \n",
    "for gruping it in groups and reduce the number of columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "sns.heatmap(df.corr(), vmin=-1, vmax=1, center= 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Impossible to see nothing so we opted to make an algorithm\n",
    "for making groups of strongly correlated variables."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "But first, we've seen that other simptomatology 4 has a big group of correlations, so we thought that\n",
    "first we shoud separate it's values to diferent cols."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "other_simptomatology_fields = ['other_simptomatology', 'other_simptomatology_3','other_simptomatology_4']\n",
    "list1 = list(map(lambda x: x == 1 , df['other_simptomatology']))\n",
    "list2 = list(map(lambda x: x == 1 , df['other_simptomatology_3']))\n",
    "list3 = list(map(lambda x: x == 1 , df['other_simptomatology_4']))\n",
    "df['dolor_toracic'] = list(map(lambda x: 1 if x else 2, \n",
    "                               map(lambda x, y, z: x or y or z, list1, list2, list3)))\n",
    "\n",
    "list1 = list(map(lambda x: x == 2 , df['other_simptomatology']))\n",
    "list2 = list(map(lambda x: x == 2 , df['other_simptomatology_3']))\n",
    "list3 = list(map(lambda x: x == 2 , df['other_simptomatology_4']))\n",
    "df['artralgies'] = list(map(lambda x: 1 if x else 2, \n",
    "                               map(lambda x, y, z: x or y or z, list1, list2, list3)))\n",
    "\n",
    "list1 = list(map(lambda x: x == 3 , df['other_simptomatology']))\n",
    "list2 = list(map(lambda x: x == 3 , df['other_simptomatology_3']))\n",
    "list3 = list(map(lambda x: x == 3 , df['other_simptomatology_4']))\n",
    "df['artritis'] = list(map(lambda x: 1 if x else 2, \n",
    "                               map(lambda x, y, z: x or y or z, list1, list2, list3)))\n",
    "\n",
    "list1 = list(map(lambda x: x == 4 , df['other_simptomatology']))\n",
    "list2 = list(map(lambda x: x == 4 , df['other_simptomatology_3']))\n",
    "list3 = list(map(lambda x: x == 4 , df['other_simptomatology_4']))\n",
    "df['mialgies'] = list(map(lambda x: 1 if x else 2, \n",
    "                               map(lambda x, y, z: x or y or z, list1, list2, list3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "for field in other_simptomatology_fields:\n",
    "    df = df.drop(field, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['sex',\n",
       " 'smokers_home',\n",
       " 'sympt_epi',\n",
       " 'home_confirmed',\n",
       " 'school_symptoms',\n",
       " 'school_confirmed',\n",
       " 'symptoms_binary',\n",
       " 'fever',\n",
       " 'highest_fever',\n",
       " 'tos',\n",
       " 'cough_first',\n",
       " 'crup',\n",
       " 'crup_first',\n",
       " 'dysphonia',\n",
       " 'disfonia_first',\n",
       " 'resp',\n",
       " 'dyspnea_first',\n",
       " 'tachypnea',\n",
       " 'tachypnea_first',\n",
       " 'ausc_resp',\n",
       " 'auscult_first',\n",
       " 'wheezing',\n",
       " 'crackles',\n",
       " 'odynophagia',\n",
       " 'odynophagia_first',\n",
       " 'nasal_congestion',\n",
       " 'nasal_first',\n",
       " 'fatiga',\n",
       " 'fatigue_first',\n",
       " 'headache',\n",
       " 'headache_first',\n",
       " 'conjuntivitis',\n",
       " 'conj_first',\n",
       " 'ocular_pain',\n",
       " 'ocular_first',\n",
       " 'gi_symptoms',\n",
       " 'gi_first',\n",
       " 'abdominal_pain',\n",
       " 'vomiting',\n",
       " 'dyarrea',\n",
       " 'dermatologic',\n",
       " 'skin_first',\n",
       " 'rash',\n",
       " 'inflam_periferic',\n",
       " 'inflam_oral',\n",
       " 'adenopathies',\n",
       " 'lymph_first',\n",
       " 'hepato',\n",
       " 'hepato_first',\n",
       " 'splenomegaly',\n",
       " 'spleno_first',\n",
       " 'hemorrhagies',\n",
       " 'hemorr_first',\n",
       " 'irritability',\n",
       " 'irritability_first',\n",
       " 'neuro',\n",
       " 'neuro_first',\n",
       " 'confusion',\n",
       " 'seizures',\n",
       " 'nuchal_stiffness',\n",
       " 'hypotonia',\n",
       " 'peripheral_paralysis',\n",
       " 'shock',\n",
       " 'shock_first',\n",
       " 'taste_smell',\n",
       " 'taste_first',\n",
       " 'smell',\n",
       " 'smell_first',\n",
       " 'sero_response',\n",
       " 'cxr2',\n",
       " 'ct2',\n",
       " 'pcr_result',\n",
       " 'antigenic_result',\n",
       " 'final_diagnosis_code',\n",
       " 'flu_binary',\n",
       " 'dolor_toracic',\n",
       " 'artralgies',\n",
       " 'artritis',\n",
       " 'mialgies']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(df.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Positive Correlations Groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import defaultdict\n",
    "def positive_correlations_group(df):\n",
    "    cor = df.corr()\n",
    "    positive_correlations = defaultdict(lambda: [])\n",
    "    for c in df.columns:\n",
    "        series = cor[[c]].drop(c)\n",
    "        series = series.dropna()\n",
    "        series = series.transpose()\n",
    "        var_corr_group = [ col for col in series.columns if series[col].item() >= 0.8]\n",
    "        positive_correlations[c] = var_corr_group\n",
    "    return positive_correlations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Negative Correlation Groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import defaultdict\n",
    "def negative_correlations_group(df):\n",
    "    cor = df.corr()\n",
    "    negative_correlations = defaultdict(lambda: [])\n",
    "    for c in df.columns:\n",
    "        series = cor[[c]].drop(c)\n",
    "        series = series.dropna()\n",
    "        series = series.transpose()\n",
    "        var_corr_group = [ col for col in series.columns if series[col].item() <= -0.8]\n",
    "        negative_correlations[c] = var_corr_group\n",
    "    return negative_correlations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we will make the groups of the variables that have a strong correlation in the same direction.\n",
    "So first we will do grups for positive correations and then for negative ones."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def purgue(el, correlations):\n",
    "    dd = defaultdict(lambda x : [])\n",
    "    for e in el:\n",
    "        dd[e] = correlations[e]\n",
    "    return dd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All this drops are in orther to simplify this variables that are represented in groups"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('tachypnea_first', axis = 1)\n",
    "df = df.drop('auscult_first', axis = 1)\n",
    "df = df.drop('smell_first', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('ocular_first', axis = 1)\n",
    "df = df.drop('adenopathies', axis = 1)\n",
    "df = df.drop('splenomegaly', axis = 1)\n",
    "df = df.drop('taste_smell', axis = 1)\n",
    "df = df.drop('confusion', axis = 1)\n",
    "df = df.drop('hypotonia', axis = 1)\n",
    "df = df.drop('cxr2', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('rash', axis = 1)\n",
    "df = df.drop('pcr_result', axis = 1)\n",
    "df = df.drop('antigenic_result', axis = 1)\n",
    "df = df.drop('smell', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('odynophagia_first', axis = 1)\n",
    "df = df.drop('dyspnea_first', axis = 1)\n",
    "df = df.drop('vomiting', axis = 1)\n",
    "df = df.drop('sero_response', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('skin_first', axis = 1)\n",
    "df = df.drop('seizures', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('cough_first', axis = 1)\n",
    "df = df.drop('fatigue_first', axis = 1)\n",
    "df = df.drop('conj_first', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(<function __main__.purgue.<locals>.<lambda>(x)>, {})"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "positive_correlations = positive_correlations_group(df)\n",
    "el1 = list(filter(lambda x: len(positive_correlations[x]) != 0, positive_correlations))\n",
    "purgue(el1, positive_correlations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.drop('inflam_oral', axis = 1)\n",
    "df = df.drop('irritability_first', axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(<function __main__.purgue.<locals>.<lambda>(x)>, {})"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "negative_correlations = negative_correlations_group(df)\n",
    "el2 = list(filter(lambda x: len(negative_correlations[x]) != 0, negative_correlations))\n",
    "purgue(el2, negative_correlations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "54"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sex</th>\n",
       "      <th>smokers_home</th>\n",
       "      <th>sympt_epi</th>\n",
       "      <th>home_confirmed</th>\n",
       "      <th>school_symptoms</th>\n",
       "      <th>school_confirmed</th>\n",
       "      <th>symptoms_binary</th>\n",
       "      <th>fever</th>\n",
       "      <th>highest_fever</th>\n",
       "      <th>tos</th>\n",
       "      <th>...</th>\n",
       "      <th>shock</th>\n",
       "      <th>shock_first</th>\n",
       "      <th>taste_first</th>\n",
       "      <th>ct2</th>\n",
       "      <th>final_diagnosis_code</th>\n",
       "      <th>flu_binary</th>\n",
       "      <th>dolor_toracic</th>\n",
       "      <th>artralgies</th>\n",
       "      <th>artritis</th>\n",
       "      <th>mialgies</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1676</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1677</th>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1678</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1679</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1680</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1681 rows × 54 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      sex  smokers_home  sympt_epi  home_confirmed  school_symptoms  \\\n",
       "0     1.0           2.0        0.0             0.0              0.0   \n",
       "1     2.0           2.0        0.0             0.0              0.0   \n",
       "2     1.0           2.0        0.0             0.0              0.0   \n",
       "3     2.0           2.0        0.0             0.0              NaN   \n",
       "4     2.0           2.0        1.0             1.0              NaN   \n",
       "...   ...           ...        ...             ...              ...   \n",
       "1676  1.0           NaN        0.0             0.0              0.0   \n",
       "1677  2.0           NaN        0.0             0.0              0.0   \n",
       "1678  1.0           2.0        0.0             0.0              0.0   \n",
       "1679  1.0           2.0        0.0             0.0              0.0   \n",
       "1680  2.0           1.0        0.0             0.0              1.0   \n",
       "\n",
       "      school_confirmed  symptoms_binary  fever  highest_fever  tos  ...  \\\n",
       "0                  0.0              1.0    1.0            1.0  1.0  ...   \n",
       "1                  0.0              1.0    1.0            2.0  1.0  ...   \n",
       "2                  0.0              1.0    1.0            2.0  2.0  ...   \n",
       "3                  NaN              1.0    2.0            NaN  1.0  ...   \n",
       "4                  NaN              1.0    1.0            3.0  1.0  ...   \n",
       "...                ...              ...    ...            ...  ...  ...   \n",
       "1676               0.0              1.0    1.0            2.0  2.0  ...   \n",
       "1677               0.0              1.0    1.0            1.0  2.0  ...   \n",
       "1678               0.0              1.0    1.0            2.0  1.0  ...   \n",
       "1679               0.0              1.0    2.0            NaN  1.0  ...   \n",
       "1680               1.0              1.0    1.0            1.0  1.0  ...   \n",
       "\n",
       "      shock  shock_first  taste_first  ct2  final_diagnosis_code  flu_binary  \\\n",
       "0       0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "1       0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "2       0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "3       0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "4       0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "...     ...          ...          ...  ...                   ...         ...   \n",
       "1676    0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "1677    0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "1678    0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "1679    0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "1680    0.0          NaN          NaN  NaN                   3.0         0.0   \n",
       "\n",
       "      dolor_toracic  artralgies  artritis  mialgies  \n",
       "0                 2           2         2         2  \n",
       "1                 2           2         2         2  \n",
       "2                 2           2         2         2  \n",
       "3                 2           2         2         2  \n",
       "4                 2           2         2         2  \n",
       "...             ...         ...       ...       ...  \n",
       "1676              2           2         2         2  \n",
       "1677              2           2         2         2  \n",
       "1678              2           2         2         2  \n",
       "1679              2           2         2         2  \n",
       "1680              2           2         2         2  \n",
       "\n",
       "[1681 rows x 54 columns]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have less variables, we do a heatmap again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "sns.heatmap(df.corr(), vmin=-1, vmax=1, center= 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Still seeing nothing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'sex': {1.0: 934, 2.0: 743, nan: 4},\n",
       " 'smokers_home': {2.0: 1074, 1.0: 451, nan: 156},\n",
       " 'sympt_epi': {0.0: 1370, 1.0: 294, nan: 17},\n",
       " 'home_confirmed': {0.0: 1411, 1.0: 119, nan: 151},\n",
       " 'school_symptoms': {0.0: 1531, 1.0: 99, nan: 51},\n",
       " 'school_confirmed': {0.0: 1415, 1.0: 84, nan: 182},\n",
       " 'symptoms_binary': {1.0: 1647, 0.0: 25, nan: 9},\n",
       " 'fever': {1.0: 983, 2.0: 662, nan: 36},\n",
       " 'highest_fever': {1.0: 285, 2.0: 535, 3.0: 162, nan: 699},\n",
       " 'tos': {1.0: 803, 2.0: 840, nan: 38},\n",
       " 'crup': {2.0: 766, 1.0: 24, 3.0: 1, nan: 890},\n",
       " 'crup_first': {0.0: 7, 1.0: 17, nan: 1657},\n",
       " 'dysphonia': {2.0: 1544, 1.0: 79, 3.0: 14, nan: 44},\n",
       " 'disfonia_first': {0.0: 10, 1.0: 69, nan: 1602},\n",
       " 'resp': {1.0: 133, 2.0: 1504, 3.0: 1, nan: 43},\n",
       " 'tachypnea': {1.0: 78, 2.0: 1550, 3.0: 10, nan: 43},\n",
       " 'ausc_resp': {2.0: 158, 1.0: 1195, nan: 328},\n",
       " 'wheezing': {1.0: 140, 2.0: 18, nan: 1523},\n",
       " 'crackles': {2.0: 107, 1.0: 46, nan: 1528},\n",
       " 'odynophagia': {2.0: 1120, 1.0: 406, 3.0: 106, nan: 49},\n",
       " 'nasal_congestion': {1.0: 849, 2.0: 779, 3.0: 4, nan: 49},\n",
       " 'nasal_first': {1.0: 794, 0.0: 50, nan: 837},\n",
       " 'fatiga': {2.0: 1225, 1.0: 362, 3.0: 47, nan: 47},\n",
       " 'headache': {2.0: 1045, 3.0: 156, 1.0: 429, nan: 51},\n",
       " 'headache_first': {1.0: 401, 0.0: 25, nan: 1255},\n",
       " 'conjuntivitis': {2.0: 1594, 1.0: 33, 3.0: 6, nan: 48},\n",
       " 'ocular_pain': {3.0: 146, 2.0: 751, 1.0: 6, nan: 778},\n",
       " 'gi_symptoms': {0.0: 1096, 1.0: 544, nan: 41},\n",
       " 'gi_first': {1.0: 493, 0.0: 42, nan: 1146},\n",
       " 'abdominal_pain': {2.0: 203, 1.0: 307, 3.0: 24, nan: 1147},\n",
       " 'dyarrea': {2.0: 234, 1.0: 308, 3.0: 1, nan: 1138},\n",
       " 'dermatologic': {0.0: 1601, 1.0: 32, nan: 48},\n",
       " 'inflam_periferic': {0.0: 31, 1.0: 1, nan: 1649},\n",
       " 'lymph_first': {1.0: 27, 0.0: 2, nan: 1652},\n",
       " 'hepato': {2.0: 1294, 3.0: 345, nan: 42},\n",
       " 'hepato_first': {nan: 1681},\n",
       " 'spleno_first': {nan: 1681},\n",
       " 'hemorrhagies': {2.0: 1496, 3.0: 131, 1.0: 3, nan: 51},\n",
       " 'hemorr_first': {1.0: 3, nan: 1678},\n",
       " 'irritability': {1.0: 37, 2.0: 632, 3.0: 10, nan: 1002},\n",
       " 'neuro': {0.0: 1628, 1.0: 10, nan: 43},\n",
       " 'neuro_first': {1.0: 8, 0.0: 2, nan: 1671},\n",
       " 'nuchal_stiffness': {0.0: 10, nan: 1671},\n",
       " 'peripheral_paralysis': {0.0: 10, nan: 1671},\n",
       " 'shock': {0.0: 1561, 1.0: 1, nan: 119},\n",
       " 'shock_first': {1.0: 1, nan: 1680},\n",
       " 'taste_first': {0.0: 12, 1.0: 17, nan: 1652},\n",
       " 'ct2': {2.0: 15, nan: 1666},\n",
       " 'final_diagnosis_code': {3.0: 1362, 1.0: 178, 2.0: 75, nan: 66},\n",
       " 'flu_binary': {0.0: 1566, 1.0: 99, nan: 16},\n",
       " 'dolor_toracic': {2: 1663, 1: 18, nan: 0},\n",
       " 'artralgies': {2: 1660, 1: 21, nan: 0},\n",
       " 'artritis': {2: 1681, nan: 0},\n",
       " 'mialgies': {2: 1636, 1: 45, nan: 0}}"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "check_values = {}\n",
    "for col in df.columns:\n",
    "    check_values[col] = get_population(df,df[col])\n",
    "check_values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sex</th>\n",
       "      <th>smokers_home</th>\n",
       "      <th>sympt_epi</th>\n",
       "      <th>home_confirmed</th>\n",
       "      <th>school_symptoms</th>\n",
       "      <th>school_confirmed</th>\n",
       "      <th>symptoms_binary</th>\n",
       "      <th>fever</th>\n",
       "      <th>highest_fever</th>\n",
       "      <th>tos</th>\n",
       "      <th>...</th>\n",
       "      <th>shock</th>\n",
       "      <th>shock_first</th>\n",
       "      <th>taste_first</th>\n",
       "      <th>ct2</th>\n",
       "      <th>flu_binary</th>\n",
       "      <th>dolor_toracic</th>\n",
       "      <th>artralgies</th>\n",
       "      <th>artritis</th>\n",
       "      <th>mialgies</th>\n",
       "      <th>final_diagnosis_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1676</th>\n",
       "      <td>1.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1677</th>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1678</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1679</th>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1680</th>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1681 rows × 54 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      sex  smokers_home  sympt_epi  home_confirmed  school_symptoms  \\\n",
       "0     1.0           2.0        0.0             0.0              0.0   \n",
       "1     2.0           2.0        0.0             0.0              0.0   \n",
       "2     1.0           2.0        0.0             0.0              0.0   \n",
       "3     2.0           2.0        0.0             0.0              NaN   \n",
       "4     2.0           2.0        1.0             1.0              NaN   \n",
       "...   ...           ...        ...             ...              ...   \n",
       "1676  1.0           NaN        0.0             0.0              0.0   \n",
       "1677  2.0           NaN        0.0             0.0              0.0   \n",
       "1678  1.0           2.0        0.0             0.0              0.0   \n",
       "1679  1.0           2.0        0.0             0.0              0.0   \n",
       "1680  2.0           1.0        0.0             0.0              1.0   \n",
       "\n",
       "      school_confirmed  symptoms_binary  fever  highest_fever  tos  ...  \\\n",
       "0                  0.0              1.0    1.0            1.0  1.0  ...   \n",
       "1                  0.0              1.0    1.0            2.0  1.0  ...   \n",
       "2                  0.0              1.0    1.0            2.0  2.0  ...   \n",
       "3                  NaN              1.0    2.0            NaN  1.0  ...   \n",
       "4                  NaN              1.0    1.0            3.0  1.0  ...   \n",
       "...                ...              ...    ...            ...  ...  ...   \n",
       "1676               0.0              1.0    1.0            2.0  2.0  ...   \n",
       "1677               0.0              1.0    1.0            1.0  2.0  ...   \n",
       "1678               0.0              1.0    1.0            2.0  1.0  ...   \n",
       "1679               0.0              1.0    2.0            NaN  1.0  ...   \n",
       "1680               1.0              1.0    1.0            1.0  1.0  ...   \n",
       "\n",
       "      shock  shock_first  taste_first  ct2  flu_binary  dolor_toracic  \\\n",
       "0       0.0          NaN          NaN  NaN         0.0              2   \n",
       "1       0.0          NaN          NaN  NaN         0.0              2   \n",
       "2       0.0          NaN          NaN  NaN         0.0              2   \n",
       "3       0.0          NaN          NaN  NaN         0.0              2   \n",
       "4       0.0          NaN          NaN  NaN         0.0              2   \n",
       "...     ...          ...          ...  ...         ...            ...   \n",
       "1676    0.0          NaN          NaN  NaN         0.0              2   \n",
       "1677    0.0          NaN          NaN  NaN         0.0              2   \n",
       "1678    0.0          NaN          NaN  NaN         0.0              2   \n",
       "1679    0.0          NaN          NaN  NaN         0.0              2   \n",
       "1680    0.0          NaN          NaN  NaN         0.0              2   \n",
       "\n",
       "      artralgies  artritis  mialgies  final_diagnosis_code  \n",
       "0              2         2         2                   3.0  \n",
       "1              2         2         2                   3.0  \n",
       "2              2         2         2                   3.0  \n",
       "3              2         2         2                   3.0  \n",
       "4              2         2         2                   3.0  \n",
       "...          ...       ...       ...                   ...  \n",
       "1676           2         2         2                   3.0  \n",
       "1677           2         2         2                   3.0  \n",
       "1678           2         2         2                   3.0  \n",
       "1679           2         2         2                   3.0  \n",
       "1680           2         2         2                   3.0  \n",
       "\n",
       "[1681 rows x 54 columns]"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def last_diagnosis_code(df):\n",
    "    diagnosis = df[['final_diagnosis_code']]\n",
    "    new_df = df.drop('final_diagnosis_code', axis=1)\n",
    "    new_df['final_diagnosis_code'] = diagnosis\n",
    "    return new_df\n",
    "df = last_diagnosis_code(df)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We're gonna drop all the variables that have a ot of NaN values \n",
    "because we thing that don't provide so much information\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "41"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for col in filter(lambda x : check_values[x][np.nan] > 1600, df.columns):\n",
    "    df = df.drop(col, axis = 1)\n",
    "len(df.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "39"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for col in filter(lambda x : check_values[x][np.nan] > 1500, df.columns):\n",
    "    df = df.drop(col, axis = 1)\n",
    "len(df.columns)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We make an algorithm to constract Bayesian Nets with random variables\n",
    "and get the best model of all the iterations.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['sympt_epi', 'odynophagia', 'artralgies', 'sex', 'ausc_resp',\n",
      "       'home_confirmed', 'headache_first', 'fever', 'resp',\n",
      "       'nasal_congestion'],\n",
      "      dtype='object')\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "General prediction result\n",
      "0.8792569659442725\n",
      "Positive prediction result\n",
      "0.36666666666666664\n",
      "Don't know prediction result\n",
      "0.0\n",
      "Negative prediction result\n",
      "0.978494623655914\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "from pomegranate import *\n",
    "import numpy\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "import networkx as nx\n",
    "\n",
    "def get_rnd_cols(df, num_vars = 10):\n",
    "    df_cols = list(df.columns)\n",
    "    columns = []\n",
    "    i = 0\n",
    "    while i < num_vars:\n",
    "        num = random.randint(0,len(df_cols)-1)\n",
    "        if df_cols[num] not in columns:\n",
    "            columns.append(df_cols[num])\n",
    "            i += 1\n",
    "    return df[columns]\n",
    "    \n",
    "\n",
    "def random_bayes(training_x, training_y):\n",
    "    training = training_x.join(training_y)\n",
    "    model = BayesianNetwork.from_samples(training, algorithm='exact')\n",
    "    return model\n",
    "\n",
    "def construct_test_set(X_test):\n",
    "    to_predict = []\n",
    "    for test in X_test.values.tolist():\n",
    "        test.append(None)\n",
    "        to_predict.append(test)\n",
    "    return to_predict\n",
    "\n",
    "def get_prediction_result(results, expected, value):\n",
    "    score = 0;\n",
    "    for i,result in enumerate(results):\n",
    "        if result[-1] == value and result[-1] == expected[i]:\n",
    "            score +=1\n",
    "    number_of_values = len([x for x in expected if x == value])\n",
    "    return score / number_of_values\n",
    "\n",
    "def get_general_prediction_result(results, expected):\n",
    "    score = 0;\n",
    "    for i,result in enumerate(results):\n",
    "        if result[-1] == expected[i]:\n",
    "            score +=1\n",
    "    number_of_values = len(expected)\n",
    "    return score / number_of_values\n",
    "\n",
    "def delete_half_negative(df, negative_cases):\n",
    "    half_negative = negative_cases / 2\n",
    "    for i in range(half_negative):\n",
    "        df.sample()\n",
    "    \n",
    "def draw_bayesian_network(graph):\n",
    "    G = nx.DiGraph()\n",
    "    for i,node in enumerate(graph):\n",
    "        G.add_node(i)\n",
    "    for i,node in enumerate(graph):\n",
    "        for edge in node:\n",
    "            G.add_edge(edge,i)\n",
    "    nx.draw(G, with_labels=True, arrowsize=10)\n",
    "    plt.show()\n",
    "\n",
    "best_result = 0\n",
    "best_pos = 0\n",
    "best_ns = 0\n",
    "best_neg=0\n",
    "df = df.dropna(subset=['final_diagnosis_code'])\n",
    "y = df['final_diagnosis_code']\n",
    "df = df.drop('final_diagnosis_code', axis = 1)\n",
    "for i in range(1000):\n",
    "    X = get_rnd_cols(df)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)\n",
    "    X_test.columns\n",
    "    model = random_bayes(X_train, y_train)\n",
    "    to_predict = construct_test_set(X_test)\n",
    "    to_predict\n",
    "    try:\n",
    "        results = model.predict(to_predict)\n",
    "    except ValueError:\n",
    "        continue\n",
    "    if get_general_prediction_result(results, y_test.values.tolist()) >= best_result:\n",
    "        best_columns = X.columns\n",
    "        best_result = get_general_prediction_result(results, y_test.values.tolist())\n",
    "        best_pos = get_prediction_result(results, y_test.values.tolist(),1)\n",
    "        best_ns = get_prediction_result(results, y_test.values.tolist(),2)\n",
    "        best_neg = get_prediction_result(results, y_test.values.tolist(),3)\n",
    "        best_model = model\n",
    "print(best_columns)\n",
    "draw_bayesian_network(best_model.structure)\n",
    "print(\"General prediction result\")\n",
    "print(best_result)\n",
    "print(\"Positive prediction result\")\n",
    "print(best_pos)\n",
    "print(\"Don't know prediction result\")\n",
    "print(best_ns)\n",
    "print(\"Negative prediction result\")\n",
    "print(best_neg)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This up here is the best model of a Bayesian Network that we got.\n",
    "\n",
    "\n",
    "It predicts the 87% of the cases. We can see that predicts the 97% of negative cases and the 36% of the positive.\n",
    "\n",
    "\n",
    "As a conclusion, I thing that the quality of the data it's not so good. Starting from the amount of NaN values that it contains, from the amount of columns in the dataset and the one that I think is the most important, ther are so much negateve cases and to feew other cases.\n",
    "\n",
    "\n",
    "I think that with a good selected set of variables, a bayesian net is a good way to go. Due to the \n",
    "charasteristics of the problem, I thing is a very good aproach. From a couple of simpthomps we infeer the \n",
    "result of the goal variable.\n",
    "\n",
    "\n",
    "But also it has limitations, and that is mainly because the algorithm that constructs the bayesian net is of exponential cost, making the number of variables that you can work with limitated. In fact, in the docs of the library that we used, advises us that we should not try to construct a net of more than 25 variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'{\\n    \"class\" : \"BayesianNetwork\",\\n    \"name\" : \"139819189142864\",\\n    \"structure\" : [\\n        [\\n            3,\\n            6,\\n            7\\n        ],\\n        [\\n            5,\\n            6,\\n            7\\n        ],\\n        [],\\n        [\\n            2\\n        ],\\n        [\\n            3,\\n            6,\\n            7\\n        ],\\n        [\\n            2,\\n            3,\\n            6,\\n            7\\n        ],\\n        [\\n            2,\\n            3\\n        ],\\n        [\\n            2,\\n            3,\\n            6\\n        ],\\n        [\\n            3,\\n            6,\\n            7\\n        ],\\n        [\\n            2,\\n            3,\\n            5,\\n            6,\\n            7\\n        ],\\n        [\\n            5,\\n            6,\\n            7\\n        ]\\n    ],\\n    \"states\" : [\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.9590163934426229\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.040983606557377046\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.98\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.02\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"0\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.7692307692307693\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.23076923076923078\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.842391304347826\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.15760869565217392\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.6666666666666666\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.7857142857142857\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.21428571428571425\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.7272727272727273\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.2727272727272727\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.9285714285714286\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.07142857142857141\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.863849765258216\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.13615023474178403\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"0.7755102040816326\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.22448979591836735\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.0\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"2.0\",\\n                                        \"0.7894736842105263\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"3.0\",\\n                                        \"0.21052631578947367\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"1.0\",\\n                                        \"0.006557377049180327\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"2.0\",\\n                                        \"0.8098360655737705\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"3.0\",\\n                                        \"0.18360655737704917\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"0.0\",\\n                                                \"0.05891980360065467\"\\n                                            ],\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\",\\n                                                \"0.9410801963993454\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            },\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"name\" : \"ConditionalProbabilityTable\",\\n                                                \"table\" : [\\n                                                    [\\n                                                        \"2.0\",\\n                                                        \"0.0\",\\n                                                        \"1.0\"\\n                                                    ]\\n                                                ],\\n                                                \"dtypes\" : [\\n                                                    \"numpy.float64\",\\n                                                    \"numpy.float64\",\\n                                                    \"float\"\\n                                                ],\\n                                                \"parents\" : [\\n                                                    {\\n                                                        \"class\" : \"Distribution\",\\n                                                        \"dtype\" : \"numpy.float64\",\\n                                                        \"name\" : \"DiscreteDistribution\",\\n                                                        \"parameters\" : [\\n                                                            {\\n                                                                \"2.0\" : 1.0\\n                                                            }\\n                                                        ],\\n                                                        \"frozen\" : false\\n                                                    }\\n                                                ]\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"1\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"dtype\" : \"numpy.float64\",\\n                \"name\" : \"DiscreteDistribution\",\\n                \"parameters\" : [\\n                    {\\n                        \"2.0\" : 1.0\\n                    }\\n                ],\\n                \"frozen\" : false\\n            },\\n            \"name\" : \"2\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"dtype\" : \"numpy.float64\",\\n                        \"name\" : \"DiscreteDistribution\",\\n                        \"parameters\" : [\\n                            {\\n                                \"2.0\" : 1.0\\n                            }\\n                        ],\\n                        \"frozen\" : false\\n                    }\\n                ]\\n            },\\n            \"name\" : \"3\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.020242914979757082\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.979757085020243\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"1.0\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"4\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.9285714285714286\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.07142857142857141\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.863849765258216\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.13615023474178403\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.7755102040816326\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.22448979591836735\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"dtype\" : \"numpy.float64\",\\n                        \"name\" : \"DiscreteDistribution\",\\n                        \"parameters\" : [\\n                            {\\n                                \"2.0\" : 1.0\\n                            }\\n                        ],\\n                        \"frozen\" : false\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"5\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"0.05891980360065467\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.9410801963993454\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"dtype\" : \"numpy.float64\",\\n                        \"name\" : \"DiscreteDistribution\",\\n                        \"parameters\" : [\\n                            {\\n                                \"2.0\" : 1.0\\n                            }\\n                        ],\\n                        \"frozen\" : false\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"6\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.7894736842105263\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.21052631578947367\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.006557377049180327\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.8098360655737705\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.18360655737704917\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"dtype\" : \"numpy.float64\",\\n                        \"name\" : \"DiscreteDistribution\",\\n                        \"parameters\" : [\\n                            {\\n                                \"2.0\" : 1.0\\n                            }\\n                        ],\\n                        \"frozen\" : false\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"7\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.9333333333333333\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.06666666666666667\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.9878542510121457\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.012145748987854251\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.9821428571428571\"\\n                    ],\\n                    [\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.01785714285714285\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"8\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.7692307692307693\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.23076923076923078\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.75\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.25\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.7704918032786885\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.2295081967213115\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.7894736842105263\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.21052631578947367\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.5\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"0.896551724137931\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.10344827586206899\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.0\",\\n                        \"0.9090909090909091\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0909090909090909\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"dtype\" : \"numpy.float64\",\\n                        \"name\" : \"DiscreteDistribution\",\\n                        \"parameters\" : [\\n                            {\\n                                \"2.0\" : 1.0\\n                            }\\n                        ],\\n                        \"frozen\" : false\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.9285714285714286\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.07142857142857141\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.863849765258216\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.13615023474178403\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"0.7755102040816326\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.22448979591836735\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.0\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"2.0\",\\n                                        \"0.7894736842105263\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"3.0\",\\n                                        \"0.21052631578947367\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"1.0\",\\n                                        \"0.006557377049180327\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"2.0\",\\n                                        \"0.8098360655737705\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"3.0\",\\n                                        \"0.18360655737704917\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"0.0\",\\n                                                \"0.05891980360065467\"\\n                                            ],\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\",\\n                                                \"0.9410801963993454\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            },\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"name\" : \"ConditionalProbabilityTable\",\\n                                                \"table\" : [\\n                                                    [\\n                                                        \"2.0\",\\n                                                        \"0.0\",\\n                                                        \"1.0\"\\n                                                    ]\\n                                                ],\\n                                                \"dtypes\" : [\\n                                                    \"numpy.float64\",\\n                                                    \"numpy.float64\",\\n                                                    \"float\"\\n                                                ],\\n                                                \"parents\" : [\\n                                                    {\\n                                                        \"class\" : \"Distribution\",\\n                                                        \"dtype\" : \"numpy.float64\",\\n                                                        \"name\" : \"DiscreteDistribution\",\\n                                                        \"parameters\" : [\\n                                                            {\\n                                                                \"2.0\" : 1.0\\n                                                            }\\n                                                        ],\\n                                                        \"frozen\" : false\\n                                                    }\\n                                                ]\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"9\",\\n            \"weight\" : 1.0\\n        },\\n        {\\n            \"class\" : \"State\",\\n            \"distribution\" : {\\n                \"class\" : \"Distribution\",\\n                \"name\" : \"ConditionalProbabilityTable\",\\n                \"table\" : [\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.07692307692307693\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"3.0\",\\n                        \"0.9230769230769231\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.25\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"3.0\",\\n                        \"0.75\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.06521739130434781\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.03260869565217392\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"3.0\",\\n                        \"0.9021739130434783\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.07894736842105263\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"3.0\",\\n                        \"0.9210526315789473\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"2.0\",\\n                        \"3.0\",\\n                        \"1.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"0.0\",\\n                        \"3.0\",\\n                        \"3.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"0.3333333333333333\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"0.03448275862068965\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"2.0\",\\n                        \"0.03448275862068965\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"2.0\",\\n                        \"3.0\",\\n                        \"0.9310344827586207\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"1.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"2.0\",\\n                        \"0.0\"\\n                    ],\\n                    [\\n                        \"2.0\",\\n                        \"1.0\",\\n                        \"3.0\",\\n                        \"3.0\",\\n                        \"1.0\"\\n                    ]\\n                ],\\n                \"dtypes\" : [\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"numpy.float64\",\\n                    \"float\"\\n                ],\\n                \"parents\" : [\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.5\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.9285714285714286\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.07142857142857141\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"1.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"1.0\",\\n                                \"0.863849765258216\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"2.0\",\\n                                \"0.13615023474178403\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"1.0\",\\n                                \"0.7755102040816326\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"2.0\",\\n                                \"0.22448979591836735\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.0\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"2.0\",\\n                                        \"0.7894736842105263\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"3.0\",\\n                                        \"0.21052631578947367\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"1.0\",\\n                                        \"0.006557377049180327\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"2.0\",\\n                                        \"0.8098360655737705\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"3.0\",\\n                                        \"0.18360655737704917\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"0.0\",\\n                                                \"0.05891980360065467\"\\n                                            ],\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\",\\n                                                \"0.9410801963993454\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            },\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"name\" : \"ConditionalProbabilityTable\",\\n                                                \"table\" : [\\n                                                    [\\n                                                        \"2.0\",\\n                                                        \"0.0\",\\n                                                        \"1.0\"\\n                                                    ]\\n                                                ],\\n                                                \"dtypes\" : [\\n                                                    \"numpy.float64\",\\n                                                    \"numpy.float64\",\\n                                                    \"float\"\\n                                                ],\\n                                                \"parents\" : [\\n                                                    {\\n                                                        \"class\" : \"Distribution\",\\n                                                        \"dtype\" : \"numpy.float64\",\\n                                                        \"name\" : \"DiscreteDistribution\",\\n                                                        \"parameters\" : [\\n                                                            {\\n                                                                \"2.0\" : 1.0\\n                                                            }\\n                                                        ],\\n                                                        \"frozen\" : false\\n                                                    }\\n                                                ]\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"0.05891980360065467\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.9410801963993454\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    },\\n                    {\\n                        \"class\" : \"Distribution\",\\n                        \"name\" : \"ConditionalProbabilityTable\",\\n                        \"table\" : [\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"0.0\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"2.0\",\\n                                \"0.7894736842105263\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"0.0\",\\n                                \"3.0\",\\n                                \"0.21052631578947367\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"1.0\",\\n                                \"0.006557377049180327\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"2.0\",\\n                                \"0.8098360655737705\"\\n                            ],\\n                            [\\n                                \"2.0\",\\n                                \"0.0\",\\n                                \"1.0\",\\n                                \"3.0\",\\n                                \"0.18360655737704917\"\\n                            ]\\n                        ],\\n                        \"dtypes\" : [\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"numpy.float64\",\\n                            \"float\"\\n                        ],\\n                        \"parents\" : [\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"dtype\" : \"numpy.float64\",\\n                                \"name\" : \"DiscreteDistribution\",\\n                                \"parameters\" : [\\n                                    {\\n                                        \"2.0\" : 1.0\\n                                    }\\n                                ],\\n                                \"frozen\" : false\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    }\\n                                ]\\n                            },\\n                            {\\n                                \"class\" : \"Distribution\",\\n                                \"name\" : \"ConditionalProbabilityTable\",\\n                                \"table\" : [\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"0.0\",\\n                                        \"0.05891980360065467\"\\n                                    ],\\n                                    [\\n                                        \"2.0\",\\n                                        \"0.0\",\\n                                        \"1.0\",\\n                                        \"0.9410801963993454\"\\n                                    ]\\n                                ],\\n                                \"dtypes\" : [\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"numpy.float64\",\\n                                    \"float\"\\n                                ],\\n                                \"parents\" : [\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"dtype\" : \"numpy.float64\",\\n                                        \"name\" : \"DiscreteDistribution\",\\n                                        \"parameters\" : [\\n                                            {\\n                                                \"2.0\" : 1.0\\n                                            }\\n                                        ],\\n                                        \"frozen\" : false\\n                                    },\\n                                    {\\n                                        \"class\" : \"Distribution\",\\n                                        \"name\" : \"ConditionalProbabilityTable\",\\n                                        \"table\" : [\\n                                            [\\n                                                \"2.0\",\\n                                                \"0.0\",\\n                                                \"1.0\"\\n                                            ]\\n                                        ],\\n                                        \"dtypes\" : [\\n                                            \"numpy.float64\",\\n                                            \"numpy.float64\",\\n                                            \"float\"\\n                                        ],\\n                                        \"parents\" : [\\n                                            {\\n                                                \"class\" : \"Distribution\",\\n                                                \"dtype\" : \"numpy.float64\",\\n                                                \"name\" : \"DiscreteDistribution\",\\n                                                \"parameters\" : [\\n                                                    {\\n                                                        \"2.0\" : 1.0\\n                                                    }\\n                                                ],\\n                                                \"frozen\" : false\\n                                            }\\n                                        ]\\n                                    }\\n                                ]\\n                            }\\n                        ]\\n                    }\\n                ]\\n            },\\n            \"name\" : \"10\",\\n            \"weight\" : 1.0\\n        }\\n    ]\\n}'"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#json = model.to_json()\n",
    "#f = open(\"./bayes_nets/model2.txt\", \"w\")\n",
    "#f.write(json)\n",
    "#json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sex</th>\n",
       "      <th>smokers_home</th>\n",
       "      <th>sympt_epi</th>\n",
       "      <th>home_confirmed</th>\n",
       "      <th>school_symptoms</th>\n",
       "      <th>school_confirmed</th>\n",
       "      <th>symptoms_binary</th>\n",
       "      <th>fever</th>\n",
       "      <th>highest_fever</th>\n",
       "      <th>tos</th>\n",
       "      <th>...</th>\n",
       "      <th>hemorrhagies</th>\n",
       "      <th>irritability</th>\n",
       "      <th>neuro</th>\n",
       "      <th>shock</th>\n",
       "      <th>flu_binary</th>\n",
       "      <th>dolor_toracic</th>\n",
       "      <th>artralgies</th>\n",
       "      <th>artritis</th>\n",
       "      <th>mialgies</th>\n",
       "      <th>final_diagnosis_code</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>184</th>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>2.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>3.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1 rows × 39 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     sex  smokers_home  sympt_epi  home_confirmed  school_symptoms  \\\n",
       "184  2.0           2.0        0.0             0.0              0.0   \n",
       "\n",
       "     school_confirmed  symptoms_binary  fever  highest_fever  tos  ...  \\\n",
       "184               0.0              1.0    1.0            2.0  1.0  ...   \n",
       "\n",
       "     hemorrhagies  irritability  neuro  shock  flu_binary  dolor_toracic  \\\n",
       "184           2.0           2.0    0.0    0.0         0.0              2   \n",
       "\n",
       "     artralgies  artritis  mialgies  final_diagnosis_code  \n",
       "184           2         2         2                   3.0  \n",
       "\n",
       "[1 rows x 39 columns]"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample = df.sample().join(y)\n",
    "sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([0.0, 2.0, 2.0, 2.0, 1.0, 0.0, 1.0, 1.0, 2.0, 1.0, 3.0],\n",
       "       dtype=object)]"
      ]
     },
     "execution_count": 146,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "column_samples = sample[best_columns]\n",
    "list_sample = column_samples.values.tolist()\n",
    "list_sample[0].append(None)\n",
    "best_model.predict(list_sample)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We try to open the model that we serialized into a file and do some predicts but it didn't work :("
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f = open(\"./bayes_nets/model2.txt\", \"r\")\n",
    "json = f.read()\n",
    "same_model = BayesianNetwork.from_json(json)\n",
    "draw_bayesian_network(same_model.structure)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "State '1' does not have key '2.0'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-150-0c76ed8c872b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0msame_model\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist_sample\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/.local/lib/python3.8/site-packages/pomegranate/BayesianNetwork.pyx\u001b[0m in \u001b[0;36mpomegranate.BayesianNetwork.BayesianNetwork.predict\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m~/.local/lib/python3.8/site-packages/pomegranate/BayesianNetwork.pyx\u001b[0m in \u001b[0;36mpomegranate.BayesianNetwork.BayesianNetwork.predict_proba\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;32m~/.local/lib/python3.8/site-packages/pomegranate/BayesianNetwork.pyx\u001b[0m in \u001b[0;36mpomegranate.BayesianNetwork._check_input\u001b[0;34m()\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: State '1' does not have key '2.0'"
     ]
    }
   ],
   "source": [
    "same_model.predict(list_sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "positive_cases = len([x for x in df.values.tolist() if x[-1] == 1])\n",
    "dk_cases = len([x for x in df.values.tolist() if x[-1] == 2])\n",
    "negative_cases = len([x for x in df.values.tolist() if x[-1] == 3])"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}