Basic_김정우_3강.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d4d6a0bb",
   "metadata": {},
   "source": [
    "# 설정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c1766b46",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 파이썬 ≥3.5 필수\n",
    "import sys\n",
    "assert sys.version_info >= (3, 5)\n",
    "\n",
    "# 사이킷런 ≥0.20 필수\n",
    "import sklearn\n",
    "assert sklearn.__version__ >= \"0.20\"\n",
    "\n",
    "# 공통 모듈 임포트\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "# 노트북 실행 결과를 동일하게 유지하기 위해\n",
    "np.random.seed(42)\n",
    "\n",
    "# 깔끔한 그래프 출력을 위해\n",
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "mpl.rc('axes', labelsize=14)\n",
    "mpl.rc('xtick', labelsize=12)\n",
    "mpl.rc('ytick', labelsize=12)\n",
    "\n",
    "# 그림을 저장할 위치\n",
    "PROJECT_ROOT_DIR = \".\"\n",
    "CHAPTER_ID = \"classification\"\n",
    "IMAGES_PATH = os.path.join(PROJECT_ROOT_DIR, \"images\", CHAPTER_ID)\n",
    "os.makedirs(IMAGES_PATH, exist_ok=True)\n",
    "\n",
    "def save_fig(fig_id, tight_layout=True, fig_extension=\"png\", resolution=300):\n",
    "    path = os.path.join(IMAGES_PATH, fig_id + \".\" + fig_extension)\n",
    "    print(\"그림 저장:\", fig_id)\n",
    "    if tight_layout:\n",
    "        plt.tight_layout()\n",
    "    plt.savefig(path, format=fig_extension, dpi=resolution)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56be7213",
   "metadata": {},
   "source": [
    "# MNIST 데이터셋 \n",
    "###### 머신러닝 분야의 hello world"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5b50b471",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['data', 'target', 'frame', 'categories', 'feature_names', 'target_names', 'DESCR', 'details', 'url'])"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.datasets import fetch_openml\n",
    "mnist = fetch_openml('mnist_784', version=1, as_frame=False)\n",
    "mnist.keys()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "693da610",
   "metadata": {},
   "source": [
    "* 'data', 'target': numpy 배열로 반환 \n",
    "* 'frame': 데이터프레임으로 반환\n",
    "* 'feature_names': 픽셀에 대한 이름\n",
    "* 'target_names': 0부터9까지의 숫자 클래스\n",
    "* 'DESCR': MNIST에 대한 상세 정보\n",
    "* 'details': openml에 대한 메타data\n",
    "* 'categories': 픽셀은 0~255 사이 정수값이기 때문에 범주값 부재\n",
    "* 'url': openml 웹사이트의 주소"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "46151e77",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 784)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = mnist[\"data\"], mnist[\"target\"]\n",
    "X.shape\n",
    "\n",
    "#data는 2차원 배열이기 때문에 대문자 X 부여"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "33569cc4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000,)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f39c245c",
   "metadata": {},
   "source": [
    "> 70000개의 샘플이 784개의 특성 픽셀(28*28)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b3347041",
   "metadata": {},
   "source": [
    "### 샘플로 그려보기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "112644b1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: some_digit_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "some_digit = X[0]\n",
    "some_digit_image = some_digit.reshape(28, 28) # 2차원 배열로 변환(numpy배열이라 reshape함수 사용 가능)\n",
    "plt.imshow(some_digit_image, cmap=mpl.cm.binary) # cmap = 'binary'흑백 반전\n",
    "plt.axis(\"off\")\n",
    "\n",
    "save_fig(\"some_digit_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "2a0316e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'5'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8b471bc",
   "metadata": {},
   "source": [
    "> 제대로 레이블된 것을 확인"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "d69a9e0a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "str"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(y[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "595f83d5",
   "metadata": {},
   "source": [
    "#### 대부분 머신러닝 알고리즘은 숫자를 기대하므로 y를 정수로 변환"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "49b24a6e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.uint8"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y = y.astype(np.uint8)\n",
    "type(y[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "b2c5fe93",
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_digit(data):\n",
    "    image = data.reshape(28, 28)\n",
    "    plt.imshow(image, cmap = mpl.cm.binary,\n",
    "               interpolation=\"nearest\")\n",
    "    plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b1e4b82",
   "metadata": {},
   "source": [
    "### 훈련집합 테스트집합 분할"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "7bbac297",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = X[:60000], X[60000:], y[:60000], y[60000:]\n",
    "# 이미 트레인셋과 테스트셋이 분리되어 60000개는 트레인셋 뒤 10000개는 테스트셋으로 가이드되어 있음"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e38fb593",
   "metadata": {},
   "source": [
    "# 이진 분류기 훈련"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "2f175507",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_5 = (y_train == 5) \n",
    "y_test_5 = (y_test == 5)\n",
    "# True와 False를 구분"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "dda4b63e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(random_state=42)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "sgd_clf = SGDClassifier(max_iter=1000, tol=1e-3, random_state=42)\n",
    "sgd_clf.fit(X_train, y_train_5)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "09836ffd",
   "metadata": {},
   "source": [
    "* SGDClassifier는 확률적 경사 하강법 분류 알고리즘으로 4강에서 상세 설명\n",
    "* SGD는 무작위성 사용하기에 결과 재현 위해선 random state 매개변수 지정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "3a20d589",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True])"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.predict([some_digit])\n",
    "# some_digit은 X[0]으로 5임\n",
    "# 사이킷런의 predict 메소드는 데이터를 2차원 배열을 가정하기에 some_digit에 []를 씌워 2차원 배열로 전달"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "a0a96109",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.95035, 0.96035, 0.9604 ])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "cross_val_score(sgd_clf, X_train, y_train_5, cv=3, scoring=\"accuracy\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4adbf318",
   "metadata": {},
   "source": [
    "# 성능 측정"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb79b551",
   "metadata": {},
   "source": [
    "### 교차 검증 사용한 정확도 측정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "c7737451",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9669\n",
      "0.91625\n",
      "0.96785\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.base import clone\n",
    "\n",
    "skfolds = StratifiedKFold(n_splits=3, random_state=42, shuffle=True)\n",
    "\n",
    "for train_index, test_index in skfolds.split(X_train, y_train_5):\n",
    "    clone_clf = clone(sgd_clf)\n",
    "    X_train_folds = X_train[train_index]\n",
    "    y_train_folds = y_train_5[train_index]\n",
    "    X_test_fold = X_train[test_index]\n",
    "    y_test_fold = y_train_5[test_index]\n",
    "\n",
    "    clone_clf.fit(X_train_folds, y_train_folds)\n",
    "    y_pred = clone_clf.predict(X_test_fold)\n",
    "    n_correct = sum(y_pred == y_test_fold)\n",
    "    print(n_correct / len(y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "712bfab6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.9669 , 0.91625, 0.96785])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(sgd_clf, X_train, y_train_5, cv=skfolds, scoring=\"accuracy\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c0fb86fe",
   "metadata": {},
   "source": [
    "> cv = skfolds 만 입력해도 위 for문 생략 가능"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "221e9e05",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator\n",
    "class Never5Classifier(BaseEstimator):\n",
    "    def fit(self, X, y=None):\n",
    "        pass\n",
    "    def predict(self, X):\n",
    "        return np.zeros((len(X), 1), dtype=bool)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "64530b25",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.91125, 0.90855, 0.90915])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "never_5_clf = Never5Classifier()\n",
    "cross_val_score(never_5_clf, X_train, y_train_5, cv=3, scoring=\"accuracy\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "299f5662",
   "metadata": {},
   "source": [
    "* 5가 아님을 판별하는 분류기 생성\n",
    "* 모두 정확도 90%를 넘지만 모든 이미지 중 단10%만이 5이기 때문에 성능이 좋다고 할 수 없음\n",
    "* 양성 클래스가 적을 때 이러한 문제가 나타남 (불균형한 데이터셋)\n",
    "* 따라서, 정확도만을 갖고 모델의 성능을 평가하기 어려움!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "756c6694",
   "metadata": {},
   "source": [
    "### 오차 행렬 (혼동 행렬)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9c0fdfcf",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import cross_val_predict\n",
    "\n",
    "y_train_pred = cross_val_predict(sgd_clf, X_train, y_train_5, cv=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b30179f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[53892,   687],\n",
       "       [ 1891,  3530]], dtype=int64)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "confusion_matrix(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33b61d52",
   "metadata": {},
   "source": [
    "![혼동행렬](https://t1.daumcdn.net/cfile/tistory/99F53C395CFE3EA919)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6de7f99e",
   "metadata": {},
   "source": [
    "뒤 N, P 먼저 읽으면: \n",
    "   * TN(음성이로 예측했는데 진짜 음성임)\n",
    "   * FP(양성으로 예측했는데 틀렸고 음성임)\n",
    "   * FN(음성으로 예측했는데 틀렸고 양성임)\n",
    "   * TP(양성으로 예측했는데 진짜 양성임)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "27c75fa6",
   "metadata": {},
   "source": [
    "![정확정밀재현](data:image/png;base64,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)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9469b792",
   "metadata": {},
   "source": [
    "#### 정밀도(Precison)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "301fabb8",
   "metadata": {},
   "source": [
    "양성으로 예측한 것으로만 계산"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed0313b8",
   "metadata": {},
   "source": [
    "#### 재현율(Recall)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1e8d2e74",
   "metadata": {},
   "source": [
    "실제 양성 데이터로만 계산"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "fea36002",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8370879772350012"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "\n",
    "precision_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "a1d9b74b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6511713705958311"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "686daa5d",
   "metadata": {},
   "source": [
    "#### f1스코어 (정밀도와 재현율의 조화평균)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "4f2d43f6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7325171197343846"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "\n",
    "f1_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5b07351d",
   "metadata": {},
   "source": [
    "### 정밀도 / 재현율 트레이드오프"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f319d853",
   "metadata": {},
   "source": [
    "분류 예측 메소드가 decision function을 호출하고 threshold 기준으로 분류\n",
    "\n",
    "임곗값이 높을수록 재현율은 낮아지고 반대로 정밀도는 높아짐"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "ff7fdad7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2164.22030239])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores = sgd_clf.decision_function([some_digit])\n",
    "y_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "a262c722",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True])"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 0\n",
    "y_some_digit_pred = (y_scores > threshold)\n",
    "y_some_digit_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "5ffc7fb9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False])"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 8000\n",
    "y_some_digit_pred = (y_scores > threshold)\n",
    "y_some_digit_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "79549875",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_scores = cross_val_predict(sgd_clf, X_train, y_train_5, cv=3,\n",
    "                             method=\"decision_function\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "1fd50975",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_train_5, y_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e295ffc1",
   "metadata": {},
   "source": [
    "#### 결정 임곗값에 대한 정밀도재현율"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "ad760e15",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: precision_recall_vs_threshold_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_precision_recall_vs_threshold(precisions, recalls, thresholds):\n",
    "    plt.plot(thresholds, precisions[:-1], \"b--\", label=\"Precision\", linewidth=2)\n",
    "    plt.plot(thresholds, recalls[:-1], \"g-\", label=\"Recall\", linewidth=2)\n",
    "    plt.legend(loc=\"center right\", fontsize=16) # Not shown in the book\n",
    "    plt.xlabel(\"Threshold\", fontsize=16)        # Not shown\n",
    "    plt.grid(True)                              # Not shown\n",
    "    plt.axis([-50000, 50000, 0, 1])             # Not shown\n",
    "\n",
    "\n",
    "\n",
    "recall_90_precision = recalls[np.argmax(precisions >= 0.90)]\n",
    "threshold_90_precision = thresholds[np.argmax(precisions >= 0.90)]\n",
    "\n",
    "\n",
    "plt.figure(figsize=(8, 4))                                                                  # Not shown\n",
    "plot_precision_recall_vs_threshold(precisions, recalls, thresholds)\n",
    "plt.plot([threshold_90_precision, threshold_90_precision], [0., 0.9], \"r:\")                 # Not shown\n",
    "plt.plot([-50000, threshold_90_precision], [0.9, 0.9], \"r:\")                                # Not shown\n",
    "plt.plot([-50000, threshold_90_precision], [recall_90_precision, recall_90_precision], \"r:\")# Not shown\n",
    "plt.plot([threshold_90_precision], [0.9], \"ro\")                                             # Not shown\n",
    "plt.plot([threshold_90_precision], [recall_90_precision], \"ro\")                             # Not shown\n",
    "save_fig(\"precision_recall_vs_threshold_plot\")                                              # Not shown\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "366f3fdc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(y_train_pred == (y_scores > 0)).all()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c301cc2f",
   "metadata": {},
   "source": [
    "#### 정밀도재현율곡선"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "03058f90",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: precision_vs_recall_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_precision_vs_recall(precisions, recalls):\n",
    "    plt.plot(recalls, precisions, \"b-\", linewidth=2)\n",
    "    plt.xlabel(\"Recall\", fontsize=16)\n",
    "    plt.ylabel(\"Precision\", fontsize=16)\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.grid(True)\n",
    "\n",
    "plt.figure(figsize=(8, 6))\n",
    "plot_precision_vs_recall(precisions, recalls)\n",
    "plt.plot([recall_90_precision, recall_90_precision], [0., 0.9], \"r:\")\n",
    "plt.plot([0.0, recall_90_precision], [0.9, 0.9], \"r:\")\n",
    "plt.plot([recall_90_precision], [0.9], \"ro\")\n",
    "save_fig(\"precision_vs_recall_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f43bebdc",
   "metadata": {},
   "source": [
    "> 그래프가 오른쪽에 붙으면 붙을수록 좋은 모델"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "4ef15638",
   "metadata": {},
   "outputs": [],
   "source": [
    "threshold_90_precision = thresholds[np.argmax(precisions >= 0.90)] #정밀도 90을 목표로 했을 때 임곗값"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "5aa96276",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3370.0194991439557"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold_90_precision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "01ae4322",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9000345901072293"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_pred_90 = (y_scores >= threshold_90_precision)\n",
    "precision_score(y_train_5, y_train_pred_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "0768dbb6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.4799852425751706"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred_90)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0b670ce0",
   "metadata": {},
   "source": [
    "### ROC 곡선"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "60b66786",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "\n",
    "fpr, tpr, thresholds = roc_curve(y_train_5, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "467c49ad",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: roc_curve_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_roc_curve(fpr, tpr, label=None):\n",
    "    plt.plot(fpr, tpr, linewidth=2, label=label)\n",
    "    plt.plot([0, 1], [0, 1], 'k--') # 대각 점선\n",
    "    plt.axis([0, 1, 0, 1])                                    # Not shown in the book\n",
    "    plt.xlabel('False Positive Rate (Fall-Out)', fontsize=16) # Not shown\n",
    "    plt.ylabel('True Positive Rate (Recall)', fontsize=16)    # Not shown\n",
    "    plt.grid(True)                                            # Not shown\n",
    "\n",
    "plt.figure(figsize=(8, 6))                                    # Not shown\n",
    "plot_roc_curve(fpr, tpr)\n",
    "fpr_90 = fpr[np.argmax(tpr >= recall_90_precision)]           # Not shown\n",
    "plt.plot([fpr_90, fpr_90], [0., recall_90_precision], \"r:\")   # Not shown\n",
    "plt.plot([0.0, fpr_90], [recall_90_precision, recall_90_precision], \"r:\")  # Not shown\n",
    "plt.plot([fpr_90], [recall_90_precision], \"ro\")               # Not shown\n",
    "save_fig(\"roc_curve_plot\")                                    # Not shown\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "24c7924d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9604938554008616"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "roc_auc_score(y_train_5, y_scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ba2cbe5b",
   "metadata": {},
   "source": [
    "## 랜덤포리스트로 훈련"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "25020bab",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "forest_clf = RandomForestClassifier(n_estimators=100, random_state=42)\n",
    "y_probas_forest = cross_val_predict(forest_clf, X_train, y_train_5, cv=3,\n",
    "                                    method=\"predict_proba\")\n",
    "# predict_proba는 예측이 아닌 예측의 확률값 반환"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "0c02330f",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_scores_forest = y_probas_forest[:, 1] # 점수 = 양성 클래스의 확률\n",
    "fpr_forest, tpr_forest, thresholds_forest = roc_curve(y_train_5,y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "11871be8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: roc_curve_comparison_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "recall_for_forest = tpr_forest[np.argmax(fpr_forest >= fpr_90)]\n",
    "\n",
    "plt.figure(figsize=(8, 6))\n",
    "plt.plot(fpr, tpr, \"b:\", linewidth=2, label=\"SGD\")\n",
    "plot_roc_curve(fpr_forest, tpr_forest, \"Random Forest\")\n",
    "plt.plot([fpr_90, fpr_90], [0., recall_90_precision], \"r:\")\n",
    "plt.plot([0.0, fpr_90], [recall_90_precision, recall_90_precision], \"r:\")\n",
    "plt.plot([fpr_90], [recall_90_precision], \"ro\")\n",
    "plt.plot([fpr_90, fpr_90], [0., recall_for_forest], \"r:\")\n",
    "plt.plot([fpr_90], [recall_for_forest], \"ro\")\n",
    "plt.grid(True)\n",
    "plt.legend(loc=\"lower right\", fontsize=16)\n",
    "save_fig(\"roc_curve_comparison_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "cadcb8a2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9983436731328145"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_train_5, y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "6e1f1bbb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9905083315756169"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_pred_forest = cross_val_predict(forest_clf, X_train, y_train_5, cv=3)\n",
    "precision_score(y_train_5, y_train_pred_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "85988041",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8662608374838591"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred_forest)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "26f0a540",
   "metadata": {},
   "source": [
    "## 다중 분류"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7cef54c0",
   "metadata": {},
   "source": [
    "* OvR, OvA: N개 모델 훈련\n",
    "* OvO: N(N-1)/2개 훈련"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "06a79fd1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=uint8)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.svm import SVC\n",
    "\n",
    "svm_clf = SVC(gamma=\"auto\", random_state=42)\n",
    "svm_clf.fit(X_train[:1000], y_train[:1000]) # y_train_5이 아니라 y_train입니다\n",
    "svm_clf.predict([some_digit])\n",
    "\n",
    "#SVC 서포트 벡터 머신은 5강에서 자세한 설명"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "247eb8aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 2.81585438,  7.09167958,  3.82972099,  0.79365551,  5.8885703 ,\n",
       "         9.29718395,  1.79862509,  8.10392157, -0.228207  ,  4.83753243]])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_digit_scores = svm_clf.decision_function([some_digit])\n",
    "some_digit_scores"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1556367d",
   "metadata": {},
   "source": [
    "> SVC는 현재 45개 이진 분류기를 훈련시켜 각 결점 점수를 얻고 가장 높은 클래스를 반환한 것"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "9220059a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(some_digit_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "181cd042",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=uint8)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "7d68d958",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "svm_clf.classes_[5]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f89fca5",
   "metadata": {},
   "source": [
    "### OvR전략 사용 강제"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "012ef9e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=uint8)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.multiclass import OneVsRestClassifier\n",
    "ovr_clf = OneVsRestClassifier(SVC(gamma=\"auto\", random_state=42))\n",
    "ovr_clf.fit(X_train[:1000], y_train[:1000])\n",
    "ovr_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "f86006af",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(ovr_clf.estimators_)\n",
    "#만든 분류기의 개수"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "fcc06eda",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([3], dtype=uint8)"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.fit(X_train, y_train)\n",
    "sgd_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "f325f07c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-31893.03095419, -34419.69069632,  -9530.63950739,\n",
       "          1823.73154031, -22320.14822878,  -1385.80478895,\n",
       "        -26188.91070951, -16147.51323997,  -4604.35491274,\n",
       "        -12050.767298  ]])"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.decision_function([some_digit])\n",
    "# SGDClassifier를 이용해 훈련해도 OvR 전략에 따라 10개 분류기 생성"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "6b10f1fd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.87365, 0.85835, 0.8689 ])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(sgd_clf, X_train, y_train, cv=3, scoring=\"accuracy\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "86bca52d",
   "metadata": {},
   "source": [
    "> cross_validate() 함수를 사용하는 편"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "942b81a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.8983, 0.891 , 0.9018])"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler\n",
    "scaler = StandardScaler()\n",
    "X_train_scaled = scaler.fit_transform(X_train.astype(np.float64))\n",
    "cross_val_score(sgd_clf, X_train_scaled, y_train, cv=3, scoring=\"accuracy\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0e3c26d",
   "metadata": {},
   "source": [
    "> (당연하지만) 전처리를 한 후 훈련을 실행하면 결과가 더 좋게 나옴"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc1f0813",
   "metadata": {},
   "source": [
    "## 에러 분석 (오차 행렬 활용)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "51a71779",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5577,    0,   22,    5,    8,   43,   36,    6,  225,    1],\n",
       "       [   0, 6400,   37,   24,    4,   44,    4,    7,  212,   10],\n",
       "       [  27,   27, 5220,   92,   73,   27,   67,   36,  378,   11],\n",
       "       [  22,   17,  117, 5227,    2,  203,   27,   40,  403,   73],\n",
       "       [  12,   14,   41,    9, 5182,   12,   34,   27,  347,  164],\n",
       "       [  27,   15,   30,  168,   53, 4444,   75,   14,  535,   60],\n",
       "       [  30,   15,   42,    3,   44,   97, 5552,    3,  131,    1],\n",
       "       [  21,   10,   51,   30,   49,   12,    3, 5684,  195,  210],\n",
       "       [  17,   63,   48,   86,    3,  126,   25,   10, 5429,   44],\n",
       "       [  25,   18,   30,   64,  118,   36,    1,  179,  371, 5107]],\n",
       "      dtype=int64)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_pred = cross_val_predict(sgd_clf, X_train_scaled, y_train, cv=3)\n",
    "conf_mx = confusion_matrix(y_train, y_train_pred)\n",
    "conf_mx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "5685d15a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: confusion_matrix_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_mx, cmap=plt.cm.gray)\n",
    "save_fig(\"confusion_matrix_plot\", tight_layout=False)\n",
    "plt.show()\n",
    "#이미지로 표현"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "bb9d10b3",
   "metadata": {},
   "outputs": [],
   "source": [
    "row_sums = conf_mx.sum(axis=1, keepdims=True)\n",
    "norm_conf_mx = conf_mx / row_sums"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "1c5bebb3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: confusion_matrix_errors_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "np.fill_diagonal(norm_conf_mx, 0)\n",
    "plt.matshow(norm_conf_mx, cmap=plt.cm.gray)\n",
    "save_fig(\"confusion_matrix_errors_plot\", tight_layout=False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e3dc1b7",
   "metadata": {},
   "source": [
    "## 다중 레이블 분류"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0161d431",
   "metadata": {},
   "source": [
    "* 여러 개의 이진 꼬리표를 출력하는 분류 시스템"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "1c384f84",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier()"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "y_train_large = (y_train >= 7)\n",
    "y_train_odd = (y_train % 2 == 1)\n",
    "y_multilabel = np.c_[y_train_large, y_train_odd]\n",
    "\n",
    "knn_clf = KNeighborsClassifier()\n",
    "knn_clf.fit(X_train, y_multilabel)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "cf6d0662",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[False,  True]])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_clf.predict([some_digit])\n",
    "# some_digit은 5이기 때문에 7보다 작으니 F, 홀수니까 T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "c680aee1",
   "metadata": {},
   "outputs": [],
   "source": [
    "#y_train_knn_pred = cross_val_predict(knn_clf, X_train, y_multilabel, cv=3)\n",
    "#f1_score(y_multilabel, y_train_knn_pred, average=\"macro\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34d72f9f",
   "metadata": {},
   "source": [
    "## 다중 출력 분류"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "72f47570",
   "metadata": {},
   "outputs": [],
   "source": [
    "noise = np.random.randint(0, 100, (len(X_train), 784))\n",
    "X_train_mod = X_train + noise\n",
    "noise = np.random.randint(0, 100, (len(X_test), 784))\n",
    "X_test_mod = X_test + noise\n",
    "y_train_mod = X_train\n",
    "y_test_mod = X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "fd7326ae",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: noisy_digit_example_plot\n"
     ]
    },
    {
     "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": [
    "some_index = 0\n",
    "plt.subplot(121); plot_digit(X_test_mod[some_index])\n",
    "plt.subplot(122); plot_digit(y_test_mod[some_index])\n",
    "save_fig(\"noisy_digit_example_plot\")\n",
    "plt.show()\n",
    "#노이즈 섞어 만든 이미지 출력"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "a64c5637",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "그림 저장: cleaned_digit_example_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "knn_clf.fit(X_train_mod, y_train_mod)\n",
    "clean_digit = knn_clf.predict([X_test_mod[some_index]])\n",
    "plot_digit(clean_digit)\n",
    "save_fig(\"cleaned_digit_example_plot\")\n",
    "#각 픽셀(레이블)에 다중 클래스(0~255)부여 ?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d48c2dd7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}