Chapter3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'COL_NAMES': ['label', 'data'],\n",
       " 'DESCR': 'mldata.org dataset: mnist-original',\n",
       " 'data': array([[0, 0, 0, ..., 0, 0, 0],\n",
       "        [0, 0, 0, ..., 0, 0, 0],\n",
       "        [0, 0, 0, ..., 0, 0, 0],\n",
       "        ..., \n",
       "        [0, 0, 0, ..., 0, 0, 0],\n",
       "        [0, 0, 0, ..., 0, 0, 0],\n",
       "        [0, 0, 0, ..., 0, 0, 0]], dtype=uint8),\n",
       " 'target': array([ 0.,  0.,  0., ...,  9.,  9.,  9.])}"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.datasets import fetch_mldata\n",
    "mnist = fetch_mldata('MNIST original')\n",
    "mnist"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((70000, 784), (70000,))"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = mnist['data'], mnist['target']\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUP8P1r5YZLcvzCQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2b9f09208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "some_digit = X[36000]\n",
    "some_digit_image = some_digit.reshape(28, 28)\n",
    "\n",
    "plt.imshow(some_digit_image, cmap=matplotlib.cm.binary, interpolation='nearest')\n",
    "plt.axis(\"off\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.0"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[36000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = X[:60000], X[60000:], y[:60000], y[60000:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "shuffle_index = np.random.permutation(60000)\n",
    "X_train, y_train = X_train[shuffle_index], y_train[shuffle_index]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train_5 = (y_train == 5)\n",
    "y_test_5 = (y_test == 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None, epsilon=0.1,\n",
       "       eta0=0.0, fit_intercept=True, l1_ratio=0.15,\n",
       "       learning_rate='optimal', loss='hinge', n_iter=5, n_jobs=1,\n",
       "       penalty='l2', power_t=0.5, random_state=42, shuffle=True, verbose=0,\n",
       "       warm_start=False)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import SGDClassifier\n",
    "\n",
    "sgd_clf = SGDClassifier(random_state=42)\n",
    "sgd_clf.fit(X_train, y_train_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True], dtype=bool)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.96735,  0.96275,  0.9674 ])"
      ]
     },
     "execution_count": 12,
     "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": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator\n",
    "\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": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.90725,  0.908  ,  0.9137 ])"
      ]
     },
     "execution_count": 14,
     "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": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "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": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[53944,   635],\n",
       "       [ 1415,  4006]])"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "confusion_matrix(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.86317603964662792"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "precision_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.73897804833056635"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7962631683561916"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "f1_score(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 190505.35207377])"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores = sgd_clf.decision_function([some_digit])\n",
    "y_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False], dtype=bool)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 200000\n",
    "y_some_digit_pred = (y_scores > threshold)\n",
    "y_some_digit_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_scores = cross_val_predict(sgd_clf, X_train, y_train_5, cv=3, method='decision_function')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precisions, recalls, thresholds = precision_recall_curve (y_train_5, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEKCAYAAAASByJ7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlO3/A92JDOyjSD80AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2b409d518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_precision_recall_vs_threshold(precisions, recalls, thresholds):\n",
    "    plt.plot(thresholds, precisions[:-1], 'b--', label=\"Precision\")\n",
    "    plt.plot(thresholds, recalls[:-1], 'g-', label=\"Recall\")\n",
    "    plt.xlabel(\"Threshold\")\n",
    "    plt.legend(loc='upper left')\n",
    "    plt.ylim([0, 1])\n",
    "\n",
    "plot_precision_recall_vs_threshold(precisions, recalls, thresholds)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train_pred_90 = (y_scores > 70000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.93282312925170063"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(y_train_5, y_train_pred_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.60708356391809626"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "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": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEKCAYAAAAMzhLIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nbtD0MV8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2b409d160>"
      ]
     },
     "metadata": {},
     "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--')\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.xlabel(\"False Positive Rate\")\n",
    "    plt.ylabel(\"True Positive Rate\")\n",
    "\n",
    "plot_roc_curve(fpr, tpr)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.96300887233758492"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "roc_auc_score(y_train_5, y_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "\n",
    "forest_clf = RandomForestClassifier(random_state=42)\n",
    "y_probas_forest = cross_val_predict(forest_clf, X_train, y_train_5, cv=3, method='predict_proba')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_scores_forest = y_probas_forest[:, 1]\n",
    "fpr_forest, tpr_forest, thresholds_forest = roc_curve(y_train_5, y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEKCAYAAAAMzhLIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n9MiwLQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2ec2a88d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(fpr, tpr, 'b:', label=\"SGD\")\n",
    "plot_roc_curve(fpr_forest, tpr_forest, \"Random Forest\")\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.99331819358199169"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_train_5, y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 5.])"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.fit(X_train, y_train)\n",
    "sgd_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-230861.5133157 , -471841.68846206, -407531.7856096 ,\n",
       "         -74888.42824017, -223754.54768902,  190505.35207377,\n",
       "        -617049.93755683, -321754.30386319, -775049.82046868,\n",
       "        -718114.9236993 ]])"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_digit_scores = sgd_clf.decision_function([some_digit])\n",
    "some_digit_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(some_digit_scores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9.])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.0"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.classes_[5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 5.])"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.multiclass import OneVsOneClassifier\n",
    "ovo_clf = OneVsOneClassifier(SGDClassifier(random_state=42))\n",
    "ovo_clf.fit(X_train, y_train)\n",
    "ovo_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "45"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(ovo_clf.estimators_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 5.])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf.fit(X_train, y_train)\n",
    "forest_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0. ,  0. ,  0. ,  0.1,  0. ,  0.9,  0. ,  0. ,  0. ,  0. ]])"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf.predict_proba([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.85822835,  0.86264313,  0.85747862])"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(sgd_clf, X_train, y_train, cv=3, scoring='accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.90931814,  0.91149557,  0.90888633])"
      ]
     },
     "execution_count": 55,
     "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')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5720,    3,   23,   11,   13,   57,   45,    8,   40,    3],\n",
       "       [   2, 6458,   48,   33,    5,   45,    5,   11,  124,   11],\n",
       "       [  52,   38, 5343,  101,   88,   20,   90,   57,  157,   12],\n",
       "       [  50,   34,  134, 5333,    3,  238,   41,   59,  143,   96],\n",
       "       [  22,   30,   34,    8, 5372,    9,   53,   30,   81,  203],\n",
       "       [  69,   42,   35,  181,   67, 4614,  116,   30,  175,   92],\n",
       "       [  29,   25,   43,    2,   50,   82, 5635,    7,   44,    1],\n",
       "       [  24,   19,   76,   26,   52,   10,    5, 5809,   16,  228],\n",
       "       [  54,  150,   72,  145,   13,  147,   54,   29, 5042,  145],\n",
       "       [  44,   31,   24,   88,  153,   36,    2,  221,   82, 5268]])"
      ]
     },
     "execution_count": 56,
     "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": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAECCAYAAADesWqHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2a724eda0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.matshow(conf_mx, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "row_sums = conf_mx.sum(axis=1, keepdims=True)\n",
    "norm_conf_mx = conf_mx / row_sums"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAECCAYAAADesWqHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAffe247h/r/jAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2b443e9e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "np.fill_diagonal(norm_conf_mx, 0)\n",
    "plt.matshow(norm_conf_mx, cmap=plt.cm.gray)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeUAAAHVCAYAAADPSuPPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSHPEbO7oLQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2a71f9c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cl_a, cl_b = 3, 5\n",
    "X_aa = X_train[(y_train == cl_a) & (y_train_pred == cl_a)]\n",
    "X_ab = X_train[(y_train == cl_a) & (y_train_pred == cl_b)]\n",
    "X_ba = X_train[(y_train == cl_b) & (y_train_pred == cl_a)]\n",
    "X_bb = X_train[(y_train == cl_b) & (y_train_pred == cl_b)]\n",
    "\n",
    "def plot_digits(instances, images_per_row=10, **options):\n",
    "    size = 28\n",
    "    images_per_row = min(len(instances), images_per_row)\n",
    "    images = [instance.reshape(size,size) for instance in instances]\n",
    "    n_rows = (len(instances) - 1) // images_per_row + 1\n",
    "    row_images = []\n",
    "    n_empty = n_rows * images_per_row - len(instances)\n",
    "    images.append(np.zeros((size, size * n_empty)))\n",
    "    for row in range(n_rows):\n",
    "        rimages = images[row * images_per_row : (row + 1) * images_per_row]\n",
    "        row_images.append(np.concatenate(rimages, axis=1))\n",
    "    image = np.concatenate(row_images, axis=0)\n",
    "    plt.imshow(image, cmap = matplotlib.cm.binary, **options)\n",
    "    plt.axis(\"off\")\n",
    "\n",
    "plt.figure(figsize=(8,8))\n",
    "plt.subplot(221); plot_digits(X_aa[:25], images_per_row=5)\n",
    "plt.subplot(222); plot_digits(X_ab[:25], images_per_row=5)\n",
    "plt.subplot(223); plot_digits(X_ba[:25], images_per_row=5)\n",
    "plt.subplot(224); plot_digits(X_bb[:25], images_per_row=5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=1, n_neighbors=5, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 63,
     "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": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[False,  True]], dtype=bool)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_train_knn_pred = cross_val_predict(knn_clf, X_train, y_train, cv=3)\n",
    "f1_score(y_train, y_train_knn_pred, average='macro')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "noise = rnd.randint(0, 100, (len(X_train), 784))\n",
    "X_train_mod = X_train + noise\n",
    "\n",
    "noise = rnd.randint(0, 100, (len(X_test), 784))\n",
    "X_test_mod = X_test + noise\n",
    "\n",
    "y_train_mod = X_train\n",
    "y_test_mod = X_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_digit(data):\n",
    "    image = data.reshape(28, 28)\n",
    "    plt.imshow(image, cmap = matplotlib.cm.binary,\n",
    "               interpolation=\"nearest\")\n",
    "    plt.axis(\"off\")\n",
    "\n",
    "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)"
   ]
  }
 ],
 "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": "3.6.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}