point_history_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import csv\n",
    "\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "RANDOM_SEED = 42"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Specify each path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Need to make further changes to that save path, so it's more iterative.\n",
    "dataset = 'model/point_history_classifier/point_history.csv'\n",
    "model_save_path = 'model/point_history_classifier/point_history_classifier2.hdf5'"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Set number of classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_CLASSES = 4"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Input length"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "TIME_STEPS = 16\n",
    "DIMENSION = 2"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Loading and splitting data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_dataset = np.loadtxt(dataset, delimiter=',', dtype='float32', usecols=list(range(1, (TIME_STEPS * DIMENSION) + 1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_dataset = np.loadtxt(dataset, delimiter=',', dtype='int32', usecols=(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X_dataset, y_dataset, train_size=0.75, random_state=RANDOM_SEED)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Building Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "use_lstm = False\n",
    "model = None\n",
    "\n",
    "if use_lstm:\n",
    "    model = tf.keras.models.Sequential([\n",
    "        tf.keras.layers.InputLayer(input_shape=(TIME_STEPS * DIMENSION, )),\n",
    "        tf.keras.layers.Reshape((TIME_STEPS, DIMENSION), input_shape=(TIME_STEPS * DIMENSION, )), \n",
    "        tf.keras.layers.Dropout(0.2),\n",
    "        tf.keras.layers.LSTM(16, input_shape=[TIME_STEPS, DIMENSION]),\n",
    "        tf.keras.layers.Dropout(0.5),\n",
    "        tf.keras.layers.Dense(10, activation='relu'),\n",
    "        tf.keras.layers.Dense(NUM_CLASSES, activation='softmax')\n",
    "    ])\n",
    "else:\n",
    "    model = tf.keras.models.Sequential([\n",
    "        tf.keras.layers.InputLayer(input_shape=(TIME_STEPS * DIMENSION, )),\n",
    "        tf.keras.layers.Dropout(0.2),\n",
    "        tf.keras.layers.Dense(24, activation='relu'),\n",
    "        tf.keras.layers.Dropout(0.5),\n",
    "        tf.keras.layers.Dense(10, activation='relu'),\n",
    "        tf.keras.layers.Dense(NUM_CLASSES, activation='softmax')\n",
    "    ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dropout (Dropout)            (None, 32)                0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 24)                792       \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 24)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                250       \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 4)                 44        \n",
      "=================================================================\n",
      "Total params: 1,086\n",
      "Trainable params: 1,086\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()  # tf.keras.utils.plot_model(model, show_shapes=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model checkpoint callback\n",
    "cp_callback = tf.keras.callbacks.ModelCheckpoint(\n",
    "    model_save_path, verbose=1, save_weights_only=False)\n",
    "# callback for early abort\n",
    "es_callback = tf.keras.callbacks.EarlyStopping(patience=20, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# model hyperparameters\n",
    "model.compile(\n",
    "    optimizer='adam',\n",
    "    loss='sparse_categorical_crossentropy',\n",
    "    metrics=['accuracy']\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model Fitting (training)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "model.fit(\n",
    "    X_train,\n",
    "    y_train,\n",
    "    epochs=1000, # might be a bit too high\n",
    "    batch_size=128,\n",
    "    validation_data=(X_test, y_test),\n",
    "    callbacks=[cp_callback, es_callback]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load modl\n",
    "model = tf.keras.models.load_model(model_save_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.9438375  0.01624594 0.01188833 0.02802826]\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "# check results\n",
    "predict_result = model.predict(np.array([X_test[0]]))\n",
    "print(np.squeeze(predict_result))\n",
    "print(np.argmax(np.squeeze(predict_result)))"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualize results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAFlCAYAAAAjyXUiAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAjKUlEQVR4nO3deZxcZZno8d/TexLWAIZsTlA2cRlAEkAGLgJCiEvABWFGwYg3eicofMbrCI4b2wyjgCNXBycOSHAEkxEZtqjEGESUJQEDZiEhbCadhLAmhCXdXfXeP/qAJSbdoft06lT178vnfFL11jl1njq29fTzPuecjpQSkiTloaHaAUiS6odJRZKUG5OKJCk3JhVJUm5MKpKk3JhUJEm5aRroHXQ+9YjnLPfBkFGHVzsESVuhq6M98nqvPL4vm3d9U27x9MWAJxVJ0lYql6odQb85/SVJyo2ViiQVRSpXO4J+M6lIUlGUTSqSpJykOqhU7KlIknJjpSJJReH0lyQpN3Uw/WVSkaSiqIPrVEwqklQUdVCp2KiXJOXGSkWSisJGvSQpL/VwnYpJRZKKwkpFkpSbOqhUbNRLknJjpSJJReF1KpKk3NTB9JdJRZKKog4a9fZUJEm5sVKRpKJw+kuSlJs6mP4yqUhSQaTk2V+SpLzUwfSXjXpJUm6sVCSpKOypSJJyUwfTXyYVSSoKb9MiScpNHVQqNuolSbmxUpGkorBRL0nKjdNfkqTclMv9X3oQEW0RcU9E3B8RiyPi3Gz8qoh4NCIWZsv+2XhExGURsSIiHoiIA3v7CFYqkjR4bAKOSiltjIhm4I6I+Fn22hdSSj95zfrHA3tly8HA5dm/W2RSkaSiGOCeSkopARuzp83ZknrYZDJwdbbdXRGxU0SMTCmt2dIGdTv9tWlTByd/6kw+eNrfM/nvPs13/vOHANx970I+MuUMTvjYZ/jS+RfT1dV9Xvg99z3AIcd+iA+dNo0PnTaNy6/8UTXDL6zjjj2SxYtu58Eld/CPX5hW7XBqhsetbwbbcUup1O+lNxHRGBELgXXAnJTS3dlLF2ZTXN+KiNZsbDSwsmLzVdnYFtVtpdLS0syVl13E0KFD6Ozq4tT/83857OB38qULLuGKb/8L4944hu98/2pu+Nkv+dD7jwPgwL9+G//+zXOrHHlxNTQ0cNm3L2TipFNYtWoNd905m5tuvpWlSx+qdmiF5nHrm0F53HKoVCJiKjC1Ymh6Smn6K09Sd+bZPyJ2Aq6PiLcB5wBrgRZgOvBF4Ly+7L/XSiUi9o2IL2bNmsuyx2/py862pYhg6NAhAHR1ddHV1UVjQwPNTU2Me+MYAA4dfyC/vO2OaoZZUyaMP4CHH36MRx/9I52dncyadQMfyBKytszj1jeD8rilcr+XlNL0lNJBFcv0ze4qpeeAecDElNKa1G0T8ANgQrZaOzC2YrMx2dgW9ZhUIuKLwI+BAO7JlgCujYizez1AVVYqlfjQadM44n2ncOj4A3j7fvtQKpVZtHQ5ALfedgdr1z316vr3L1rKB0/7ez7z+a+w4pHHqxV2YY0avTsrV61+9fmq9jWMGrV7FSOqDR63vvG45S8idssqFCJiCPAe4MGIGJmNBXACsCjb5Ebg1OwssEOA9T31U6D36a/TgbemlDpfE9ilwGLgoi0E/mr59e+XXMCnTj2ll90MjMbGRq6b8V02PL+RM885nxWPPs43zzubb1w2nY7OTt414UAaGrrz6n77vJk5181g6NAh3P67e/jcOecxe+YVVYlb0iA18Bc/jgRmREQj3UXFrJTSzRHxq4jYje6iYSHwmWz92cAkYAXwIjCltx30llTKwCjgtb+2j8xe26ys3JoO0PnUIz2dWbBN7LD9dkw48B3ccdcCpvzth7n68osB+O3d9/L4yu5Kbrthw15d/4h3TeCCS77Ls8+tZ+eddqxKzEW0un0tY8eMevX5mNEjWb16bRUjqg0et74ZlMdtgC9+TCk9ABywmfGjtrB+Al7XGRK99VTOAuZGxM8iYnq2/ByYC5z5ena0rT3z7HNseL77zLmXN23izvm/Z4+/GsvTzz4HQEdHB1f+6L856YRJADz19DN0Hz/4w5JllFNipx13qErsRTV/wUL23HMPxo0bS3NzMyedNJmbbr612mEVnsetbwblcRvgix+3hR4rlZTSzyNib7qbNq+cRtYOzE8F/2PKTz79LP90wcWUymVSOXHcUYdz5GEHc/F3/pNf/+4eUrnMR098Lwe/c38Abp13BzOvv4XGpkbaWlr45rln0z29qFeUSiXOPOvLzL7lGhobGrhqxkyWLFle7bAKz+PWN4PyuNXBbVrild/OB0oRpr9q0ZBRh1c7BElboaujPbffPl/6xXf6/X055LgzqvrbcN1epyJJNacA01f9ZVKRpKIwqUiSclMHPZW6vfeXJGnbs1KRpKJw+kuSlJs6mP4yqUhSUVipSJJyUweVio16SVJurFQkqSic/pIk5cakIknKzQDfi3FbMKlIUlHUQaVio16SlBsrFUkqijqoVEwqklQUdXCdiklFkoqiDioVeyqSpNxYqUhSUXhKsSQpN3Uw/WVSkaSiMKlIknJTB2d/2aiXJOXGSkWSCiKVbdRLkvJiT0WSlJs66KmYVCSpKOpg+stGvSQpN1YqklQU9lQkSbmpg6Ti9JckFUVK/V96EBFtEXFPRNwfEYsj4txsfI+IuDsiVkTEzIhoycZbs+crstfH9fYRTCqSNHhsAo5KKf01sD8wMSIOAf4V+FZKaU/gWeD0bP3TgWez8W9l6/XIpCJJRVEu93/pQeq2MXvanC0JOAr4STY+Azghezw5e072+tERET3tw6QiSUVRTv1eImJqRCyoWKZW7iIiGiNiIbAOmAM8DDyXUurKVlkFjM4ejwZWAmSvrwd26ekj2KiXpKLI4eLHlNJ0YHoPr5eA/SNiJ+B6YN9+77SCSUWSimIbXvyYUnouIuYBhwI7RURTVo2MAdqz1dqBscCqiGgCdgSe7ul9BzypDBl1+EDvoi6t/+q7qx1CTdrxvHnVDqEm9ThJrroREbsBnVlCGQK8h+7m+zzgw8CPgdOAG7JNbsye35m9/quUej7FzEpFkgoiDfx1KiOBGRHRSHdPfVZK6eaIWAL8OCIuAH4PXJGtfwXww4hYATwDnNzbDkwqklQUAzz9lVJ6ADhgM+OPABM2M/4y8JHXsw+TiiQVRR3cpdhTiiVJubFSkaSiqINb35tUJKko6uCGkiYVSSoKKxVJUm5s1EuS9CdWKpJUFE5/SZLysg2uqB9wJhVJKgorFUlSbuogqdiolyTlxkpFkoqiDk4pNqlIUlHUwfSXSUWSCiLVQVKxpyJJyo2ViiQVRR1UKiYVSSoKL36UJOXGSkWSlJs6SCo26iVJubFSkaSCSKn2KxWTiiQVRR1Mf5lUJKkoTCqSpLx4Rb0kSRWsVCSpKOqgUjGpSFJR1P4F9SYVSSoKeyqSJFWwUpGkoqiDSsWkIklFUQc9Fae/JKkgUjn1e+lJRIyNiHkRsSQiFkfEmdn41yOiPSIWZsukim3OiYgVEbEsIo7r7TNYqUhSUQx8pdIFfD6ldF9EbA/cGxFzste+lVK6uHLliNgPOBl4KzAK+GVE7J1SKm1pB4M2qRx37JFceul5NDY0cOUPruUb3/xutUMqjNhhOK0f+AwxbEcg0XnfPLrm/4KGN7yRlklTiJY2ys89yab/uRw6XqJhj7fRctRHicYmUqmLjrnXUn5sSbU/RqH48/b6tba2Mu9X19Ha2kpjUyM//ektnHfeJdUOq6allNYAa7LHz0fEUmB0D5tMBn6cUtoEPBoRK4AJwJ1b2mBQJpWGhgYu+/aFTJx0CqtWreGuO2dz0823snTpQ9UOrRjKZTp+eQ3ltY9BSxtDTj+f0qN/oOV9n+oe/+ODNP31ETQf+l46f/0TePF5Ns28hLTxOWK3MbSd8o+8dNnnqv0pCsOft77ZtGkT7zn2JF544UWampr49W3X84ufz+Pue+6rdmgDJo9TiiNiKjC1Ymh6Smn6ZtYbBxwA3A0cBpwREacCC+iuZp6lO+HcVbHZKnpOQoOzpzJh/AE8/PBjPProH+ns7GTWrBv4wPt7nSocNNLG57oTCkDHy5SfWk1sP5yG4btT/uODAJQeXUTTvuMBKD/xOGnjc93bPrmKaG6BxkH5+8pm+fPWdy+88CIAzc1NNDc318Wt4XtU7v+SUpqeUjqoYtlcQtkOuA44K6W0AbgceDOwP92VTJ9Lwj4nlYiY0tdtq23U6N1ZuWr1q89Xta9h1KjdqxhRccWOu9Kw+19Rbn+Y8pOraNz7nQA0vuVgYofhf7F+477juxNSqWsbR1pc/rz1XUNDAwvm38rq9gf45dzbuWf+76sd0oBK5f4vvYmIZroTyo9SSj8FSCk9kVIqpZTKwPfpnuICaAfGVmw+Jhvbov5UKuf2EPTUiFgQEQvK5Rf6sQtVVXMrrR8+k45b/ws6XmLTzd+n+aBjaDv9fKKl7S8SR+w6mpajT2bT7CurFLDqTblc5qDxxzJuj4MYf9ABvPWt+1Q7pIGVQ6XSk4gI4ApgaUrp0orxkRWrnQgsyh7fCJwcEa0RsQewF3BPT/vocY4iIh7Y0kvAiC1tl5Vb0wGaWkYXrl5d3b6WsWNGvfp8zOiRrF69tooRFVBDI60fPpOuRb+jtGwBAOnpNbx8zb8CEMN3p3HP/V9dPbYfTttHzmLTDd8jPbuuGhEXlj9v/bd+/QZu+/VvOfbYI1m8eFm1w6llhwEfB/4QEQuzsS8Bp0TE/kACHgM+DZBSWhwRs4AldJ85Nq2nM7+g90b9COA44NnXjAfwu639FEUzf8FC9txzD8aNG0t7+1pOOmkyHz91WrXDKpSW932K9NRquu7+2Z8Gh+4AL24Agua/mUzXfXO7x1uH0nry5+n41UzKq2w+v5Y/b32z667D6ezsYv36DbS1tXHM0UfwzYv/vdphDaitmb7q1/undAfd39+vNbuHbS4ELtzaffSWVG4GtkspLXztCxFx29bupGhKpRJnnvVlZt9yDY0NDVw1YyZLliyvdliF0TB2b5rfcTjlJ/5I26e6f5Y6580ihu9O80HHAND14AK67r8dgObx76Fh5xE0H34izYefCNBd0by4oTofoGD8eeubkSNHcOUV/0ZjYwPR0MBPfnITs2f/stphDaw6uKI+BvpsiiJOf9WC9V99d7VDqEk7njev2iHUpM396qqt09nRntvhe/I9/6vf35e7zfl1Vf/nHJSnFEuSBoYXE0hSQQx0T2VbMKlIUkGYVCRJ+Um1390yqUhSQdRDpWKjXpKUGysVSSqIVHb6S5KUk3qY/jKpSFJBJBv1kqS81EOlYqNekpQbKxVJKggb9ZKk3NTDX0s2qUhSQdRDpWJPRZKUGysVSSqIeqhUTCqSVBD2VCRJubFSkSTlph6uqLdRL0nKjZWKJBVEPdymxaQiSQVRroPpL5OKJBVEPfRUTCqSVBD1cPaXjXpJUm6sVCSpILz4UZKUm3qY/jKpSFJB1MPZX/ZUJEm5sVKRpIKoh1OKrVQkqSBS6v/Sk4gYGxHzImJJRCyOiDOz8eERMSciHsr+3Tkbj4i4LCJWRMQDEXFgb5/BpCJJBVFO0e+lF13A51NK+wGHANMiYj/gbGBuSmkvYG72HOB4YK9smQpc3tsOTCqSVBApRb+Xnt8/rUkp3Zc9fh5YCowGJgMzstVmACdkjycDV6dudwE7RcTInvZhUpGkQSgixgEHAHcDI1JKa7KX1gIjssejgZUVm63KxrbIpCJJBZFHTyUipkbEgopl6mv3ExHbAdcBZ6WUNvx5DCkBfb4M07O/CmrH8+ZVO4SatOHC46odQk3a5atzqx2CyOc6lZTSdGD6ll6PiGa6E8qPUko/zYafiIiRKaU12fTWumy8HRhbsfmYbGyLrFQkqSAGuqcSEQFcASxNKV1a8dKNwGnZ49OAGyrGT83OAjsEWF8xTbZZViqSVBDb4Ir6w4CPA3+IiIXZ2JeAi4BZEXE68DhwUvbabGASsAJ4EZjS2w5MKpI0SKSU7gC2lLmO3sz6CZj2evZhUpGkgqiDmxSbVCSpKOrhhpImFUkqCO/9JUlSBSsVSSqIcrUDyIFJRZIKIm3xxKzaYVKRpIIo18HpXyYVSSqIch1UKjbqJUm5sVKRpIKwpyJJyo1nf0mSclMPlYo9FUlSbqxUJKkgnP6SJOXGpCJJyk099FRMKpJUEOXazyk26iVJ+bFSkaSCqIfbtJhUJKkg6uB+kiYVSSoKz/6SJOWmHLU//WWjXpKUGysVSSoIeyqSpNzYU5Ek5caLHyVJqmClIkkF4cWPkqTc2KiXJOWmHnoqJhVJKoh6OPvLRr0kKTdWKpJUEPXQUxm0lcpxxx7J4kW38+CSO/jHL0yrdjg1w+O2ZbH9zrR+9Au0TbmAtinn03TgMd3ju42l9e++RNsnzqP1xM9BS1v3Bg2NtBx/Om2fOI+2T15A08GTqhh9MU2bNoUFC27l3nvncMYZn6x2OAOuHP1fehMRV0bEuohYVDH29Yhoj4iF2TKp4rVzImJFRCyLiON6e/9BWak0NDRw2bcvZOKkU1i1ag133Tmbm26+laVLH6p2aIXmcetZKpfpmDeTtO6P0NxG26lfpfT4ElqO+wSdt82kvGo5jW/7G5rHH0/nb6+ncZ+DoLGJl6/6KjS10PbJCygtvZu04elqf5RC2G+/vZky5RQOP/wDdHR0cuONVzN79lweeeTxaoc2YLZRT+Uq4DvA1a8Z/1ZK6eLKgYjYDzgZeCswCvhlROydUipt6c17rVQiYt+IODoitnvN+MSti794Jow/gIcffoxHH/0jnZ2dzJp1Ax94f68JeNDzuPXihfXdCQWg82XKT68httuJhuEjKK9aDkD58cU07v3O7nUSRHMrRAM0NUOpi9TxcpWCL559992T+fMX8tJLL1MqlfjNb+7mhBNq9mtnq5RzWHqTUrodeGYrQ5oM/DiltCml9CiwApjQ0wY9JpWI+BxwA/BZYFFETK54+Z+3MqjCGTV6d1auWv3q81Xtaxg1avcqRlQbPG5bL3bYhYYRb6S85hHKT62mcc8DAGjcZzyxw3AASssXkDo3MeTvv8WQT19M5/xfwMsvVDPsQlm8eDmHHTae4cN3YsiQNiZOfDdjxoyqdliFFxFTI2JBxTJ1Kzc9IyIeyKbHds7GRgMrK9ZZlY1tUW/TX/8beGdKaWNEjAN+EhHjUkrfhi1f+pl9iKkA0bgjDQ3DetmNVEeaW2mdPI3OX10LHS/T8fMraTn6b2k+9P10PbwQSl0ANIzcA1KZly7/B2gbStsp51B+fAlp/ZPVjb8gli1bwSWXfI+bbvovXnzxRe6/fzGl0hZnXepCyuE6lZTSdGD669zscuB8us8VOB+4BOhTE6u3pNKQUtoIkFJ6LCKOpDux/BU9JJXKD9XUMrpwJzSsbl/L2IrfeMaMHsnq1WurGFFt8LhthYZGWidPo2vpXZQeug+A9MxaNv33pQDEziNofNM7AGh8yyGUHl0E5RK8+Dzl9odo2H0cJZPKq2bMmMmMGTMBOPfcL9DeXt8/b9W6TiWl9MQrjyPi+8DN2dN2YGzFqmOysS3qrafyRETsX7HjjcD7gF2Bt299yMUyf8FC9txzD8aNG0tzczMnnTSZm26+tdphFZ7HrXctE6dQfnoNXQsqjsvQ7bMH0V2tLLwNgLThaRrf+Jbul5pbaBj5ZsrPrNmm8RbdbrvtAsDYsaOYPHkiM2feUOWIBta26KlsTkSMrHh6IvDKmWE3AidHRGtE7AHsBdzT03v1VqmcCnRVDqSUuoBTI+I/XlfUBVIqlTjzrC8z+5ZraGxo4KoZM1myZHm1wyo8j1vPGkbvRdNb30X5yZU0nvZ1ADpuv46GnUfQdMBRAJQeuo/SojsA6Pr9r2g5/pO0TTkfCLoW3UF6clWVoi+ma6/9HsOH70xnZydnnfVV1q/fUO2Qal5EXAscCewaEauArwFHZgVEAh4DPg2QUlocEbOAJXTngmk9nfkFECkN7OxUEae/VL82XOjZaH2xy1fnVjuEmvXSS4/ndseu/zf2Y/3+vvzsyv+q6h3EBuV1KpJURN5QUpKUm3q4oaRJRZIKoh6SyqC995ckKX9WKpJUEPVwVpNJRZIKwka9JCk39dBTMalIUkHUw/SXjXpJUm6sVCSpIMp1UKuYVCSpIOypSJJyU/t1ij0VSVKOrFQkqSCc/pIk5caLHyVJufHsL0lSbmo/pdiolyTlyEpFkgrCRr0kKTf2VCRJuan9lGJSkaTCqIfpLxv1kqTcWKlIUkHYU5Ek5ab2U4pJRZIKw56KJEkVrFQkqSBSHUyAmVQkqSDqYfrLpCJJBeHZX5Kk3NR+SrFRL0nKkUlFkgqiTOr30puIuDIi1kXEooqx4RExJyIeyv7dORuPiLgsIlZExAMRcWBv729SkaSCKOewbIWrgImvGTsbmJtS2guYmz0HOB7YK1umApf39uYmFUkqiJTDf73uI6XbgWdeMzwZmJE9ngGcUDF+dep2F7BTRIzs6f1NKpJUEHlUKhExNSIWVCxTt2LXI1JKa7LHa4ER2ePRwMqK9VZlY1s04Gd/NUQM9C7q0o5tw6odQk3a+Stzqh1CTXr23GOqHYJyklKaDkzvx/YpIvp8IpqnFEtSQVTxivonImJkSmlNNr21LhtvB8ZWrDcmG9sip78kqSC2UaN+c24ETssenwbcUDF+anYW2CHA+oppss2yUpGkgiinga9UIuJa4Ehg14hYBXwNuAiYFRGnA48DJ2WrzwYmASuAF4Epvb2/SUWSBpGU0ilbeOnozaybgGmv5/1NKpJUEPVwmxaTiiQVhDeUlCTlxr+nIknKTT38PRVPKZYk5cZKRZIKwp6KJCk39lQkSbmph56KSUWSCiJtgyvqB5qNeklSbqxUJKkgbNRLknJjT0WSlJt6OPvLnookKTdWKpJUEPZUJEm5qYdTik0qklQQNuolSbmxUS9JUgUrFUkqCBv1kqTc2KiXJOWmHioVeyqSpNxYqUhSQdTD2V8mFUkqiLI9FUlSXmo/pZhUJKkwbNRLklTBSkWSCqIeKhWTiiQVhBc/SpJyY6UiScqN16nUuIaGBu66czbtq9dy4omfqHY4hTRq9O5893vfYLc37EJKiR9eNYvp37uaD5wwkS+cfQZ77/Nmjj3qI9z/+0XVDrXQli37HRuff4FSqURXV4l3HfbeaodUGLH9cFre+yli2A4AdC38NV33ziF2G0vLcacSLW2k9U+x6ab/gI6XadzvEJonHP+n7d8whpev+jpp3cpqfYSaEhGPAc8DJaArpXRQRAwHZgLjgMeAk1JKz/bl/Qd1UvnsZ0/nwQdXsP0O21U7lMIqdZX42pcv4oH7lzBsu2HM/fV13DbvtyxdspxPfOyzXPJv51Y7xJpx7HEn8fTTffr/aV1L5RId82aSnngcWtpoO+1rlB5bTMvxU+icN5PyymU0vv1wmg8+ns7fXE9pyV2UltwFQOw6htYPfrZuEso27Km8O6X0VMXzs4G5KaWLIuLs7PkX+/LGvZ5SHBETImJ89ni/iPiHiJjUl50VyejRIzn++KO58gfXVDuUQnviiSd54P4lALyw8QWWL3uEkaNG8NDyR3h4xaNVjk514YX13QkFoONlyk+vIbbfiYbhIyivXAZA+bHFNO79zr/YtGm/gyktvXtbRjugyqR+L300GZiRPZ4BnNDXN+oxqUTE14DLgMsj4l+A7wDDgLMj4p/6utMiuOTir3POORdSLtf+HOa2MvaNo3n7O97CvQvur3YotSclbrn5R9z5u1s4/fS/rXY0hRU77ELDiDdSXv0I5adW07jXAQA07nsQsf3wv1i/cd8JdNVRUkkp9XvZmt0At0bEvRExNRsbkVJakz1eC4zo62fobfrrw8D+QGu2ozEppQ0RcTFwN3Dh5jbKAp0K0Ni4Ew2Nw/oa34CYNOlo1j35FL///R844ohDqx1OTRg2bCg/+OFlfPmcf2bj8y9UO5ya8+6jPsTq1WvZbbddmH3LNSxb9jB33FE/X4a5aG6l9cQz6Jx7LXS8TMfsK2g55u9oftcH6FqxEMqlP1u9YeSboKuD9FR7deItqMrv38z0lNL0iud/k1Jqj4g3AHMi4sHK7VNKKSL6/Nt2b0mlK6VUAl6MiIdTShuynb4UEeUtbZR9gOkALa1jClcKvOvQ8bzvvccy8bijaGtrZYcdtueqH1zGJ6Z8rtqhFVJTUxM/+OFl/GTWTdxy05xqh1OTVq9eC8CTTz7NDTf+nPEH7W9SqdTQSOuJZ9C15E5Ky+8FID2zlk2zLgEgdh5B45ve8WebNL5lAl1Zb6Ve5HFKceX37xZeb8/+XRcR1wMTgCciYmRKaU1EjATW9XX/vfVUOiJiaPb41QnNiNgR2GJSKbovf+Ui3vTm8ey9z6F87OPTmHfbb00oPfi371zI8mWP8L3vXlXtUGrS0KFD2G67Ya8+PuboI1i8eFmVoyqWluOnUH56NV3zb/3T4NDtswdB87veT9fC2yq2CBr3nUBp6T3bMMqBl3L4rycRMSwitn/lMXAssAi4ETgtW+004Ia+fobeKpUjUkqbAFJKlUmkuSIA1bGDD3knHz3lBBYvWsa83/wPABeedyktrS38yze+wi67DueaWf/B4j8s5aQPfqq6wRbUiBG7MWvm9wFoamrkxzNv4NY5t1U3qAJpGL0XTW87jPK6lTR+ovtswo7br6Nh5xE0HXgUAKXl91L6w2/+tM3YvUnPP0Na/2RVYh4o2+DW9yOA6yMCur//r0kp/Twi5gOzIuJ04HHgpL7uIAb6FLYiTn/Vgh3bitWHqhUbNr1Y7RBq0rPnHlPtEGrW0C/+IPJ6r7eOOLjf35eLn7g7t3j6wrsUS5JyM6gvfpSkIvEvP0qScuO9vyRJubFSkSTlph4qFRv1kqTcWKlIUkE4/SVJyk09TH+ZVCSpIP78xiW1yZ6KJCk3ViqSVBB53KW42kwqklQQ2/DPCQ8Yk4okFYSViiQpN/VQqdiolyTlxkpFkgrCix8lSbnx4kdJUm7qoadiUpGkgqiHs79s1EuScmOlIkkF4fSXJCk3nv0lScpNPVQq9lQkSbmxUpGkgqiHs79MKpJUEPUw/WVSkaSCsFEvScpNPdymxUa9JCk3ViqSVBBOf0mScmOjXpKUG3sqkqTcpJT6vfQmIiZGxLKIWBERZ+f9GUwqkjRIREQj8F3geGA/4JSI2C/PfTj9JUkFsQ16KhOAFSmlRwAi4sfAZGBJXjuwUpGkgkg5LL0YDayseL4qG8vNgFcqHZtWxUDvo68iYmpKaXq146g1Hre+89j1zWA5bl0d7f3+voyIqcDUiqHp2/LYDfZKZWrvq2gzPG5957HrG4/bVkopTU8pHVSxVCaUdmBsxfMx2VhuBntSkaTBZD6wV0TsEREtwMnAjXnuwEa9JA0SKaWuiDgD+AXQCFyZUlqc5z4Ge1Kp+znaAeJx6zuPXd943HKSUpoNzB6o9496uC2AJKkY7KlIknIzaJPKQN+qoB5FxJURsS4iFlU7lloSEWMjYl5ELImIxRFxZrVjqgUR0RYR90TE/dlxO7faMal3g3L6K7tVwXLgPXRf/DMfOCWllNtVpfUoIo4ANgJXp5TeVu14akVEjARGppTui4jtgXuBE/x561lEBDAspbQxIpqBO4AzU0p3VTk09WCwViqv3qogpdQBvHKrAvUgpXQ78Ey146g1KaU1KaX7ssfPA0vJ+SrmepS6bcyeNmfL4PstuMYM1qQy4LcqkDYnIsYBBwB3VzmUmhARjRGxEFgHzEkpedwKbrAmFWmbi4jtgOuAs1JKG6odTy1IKZVSSvvTfeX3hIhw2rXgBmtSGfBbFUiVsp7AdcCPUko/rXY8tSal9BwwD5hY5VDUi8GaVAb8VgXSK7KG8xXA0pTSpdWOp1ZExG4RsVP2eAjdJ9Y8WNWg1KtBmVRSSl3AK7cqWArMyvtWBfUoIq4F7gT2iYhVEXF6tWOqEYcBHweOioiF2TKp2kHVgJHAvIh4gO5fBOeklG6uckzqxaA8pViSNDAGZaUiSRoYJhVJUm5MKpKk3JhUJEm5MalIknJjUpEk5cakIknKjUlFkpSb/w8J9yzUaL6/qwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 504x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification Report\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.99      1.00      0.99       395\n",
      "           1       0.93      0.99      0.96       295\n",
      "           2       0.98      0.97      0.98       307\n",
      "           3       0.96      0.91      0.93       327\n",
      "\n",
      "    accuracy                           0.97      1324\n",
      "   macro avg       0.97      0.97      0.97      1324\n",
      "weighted avg       0.97      0.97      0.97      1324\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix, classification_report\n",
    "\n",
    "def print_confusion_matrix(y_true, y_pred, report=True):\n",
    "    labels = sorted(list(set(y_true)))\n",
    "    cmx_data = confusion_matrix(y_true, y_pred, labels=labels)\n",
    "    \n",
    "    df_cmx = pd.DataFrame(cmx_data, index=labels, columns=labels)\n",
    " \n",
    "    fig, ax = plt.subplots(figsize=(7, 6))\n",
    "    sns.heatmap(df_cmx, annot=True, fmt='g' ,square=False)\n",
    "    ax.set_ylim(len(set(y_true)), 0)\n",
    "    plt.show()\n",
    "    \n",
    "    if report:\n",
    "        print('Classification Report')\n",
    "        print(classification_report(y_test, y_pred))\n",
    "\n",
    "Y_pred = model.predict(X_test)\n",
    "y_pred = np.argmax(Y_pred, axis=1)\n",
    "\n",
    "print_confusion_matrix(y_test, y_pred)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Conver model to Tensorflow-Lite"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save as inference-only model\n",
    "model.save(model_save_path, include_optimizer=False)\n",
    "model = tf.keras.models.load_model(model_save_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "tflite_save_path = 'model/point_history_classifier/point_history_classifier2.tflite'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# transform model (quantization)\n",
    "converter = tf.lite.TFLiteConverter.from_keras_model(model)  # converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_path)\n",
    "converter.optimizations = [tf.lite.Optimize.DEFAULT]\n",
    "tflite_quantized_model = converter.convert()\n",
    "\n",
    "open(tflite_save_path, 'wb').write(tflite_quantized_model)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Check results again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "interpreter = tf.lite.Interpreter(model_path=tflite_save_path)\n",
    "interpreter.allocate_tensors()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[{'name': 'input_1', 'index': 0, 'shape': array([ 1, 32]), 'shape_signature': array([-1, 32]), 'dtype': <class 'numpy.float32'>, 'quantization': (0.0, 0), 'quantization_parameters': {'scales': array([], dtype=float32), 'zero_points': array([], dtype=int32), 'quantized_dimension': 0}, 'sparsity_parameters': {}}]\n"
     ]
    }
   ],
   "source": [
    "# get input and output of tensors\n",
    "input_details = interpreter.get_input_details()\n",
    "output_details = interpreter.get_output_details()\n",
    "print(input_details)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "interpreter.set_tensor(input_details[0]['index'], np.array([X_test[0]]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Inference implementation\n",
    "interpreter.invoke()\n",
    "tflite_results = interpreter.get_tensor(output_details[0]['index'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.9579909  0.01342559 0.00907356 0.01950999]\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "print(np.squeeze(tflite_results))\n",
    "print(np.argmax(np.squeeze(tflite_results)))"
   ]
  }
 ],
 "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}