3_DeteccionClasificacionGH/03S_5_clasificadores.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Clasificadores - conceptos básicos**\n",
    "\n",
    "*Códigos referencia del libro Hands-On Machine Learning with Scikit-Learn and Tensor Flow 2017 (Cap 3) -- Aurélien Géron"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Configuraciones básicas"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Preparar módulos de Python:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To support both python 2 and python 3\n",
    "from __future__ import division, print_function, unicode_literals\n",
    "\n",
    "# Common imports\n",
    "import numpy as np\n",
    "import os\n",
    "\n",
    "# to make this notebook's output stable across runs\n",
    "np.random.seed(42)\n",
    "\n",
    "# To plot pretty figures\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",
    "# Where to save the figures\n",
    "PROJECT_ROOT_DIR = \".\"\n",
    "CHAPTER_ID = \"classification\"\n",
    "\n",
    "def save_fig(fig_id, tight_layout=True):\n",
    "    path = os.path.join(PROJECT_ROOT_DIR, \"imagesAM\", CHAPTER_ID, fig_id + \".png\")\n",
    "    print(\"Saving figure\", fig_id)\n",
    "    if tight_layout:\n",
    "        plt.tight_layout()\n",
    "    plt.savefig(path, format='png', dpi=300)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MNIST : \"Hello world\" de clasificación "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se ordenan los datos para asegurar repetibilidad de los experimentos. Note que se ordenan por separado el conjunto de entrenamiento (primeras 60000 imágenes) y el conjunto de evaluación (las restantes 10000)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def sort_by_target(mnist):\n",
    "    reorder_train = np.array(sorted([(target, i) for i, target in enumerate(mnist.target[:60000])]))[:, 1]\n",
    "    reorder_test = np.array(sorted([(target, i) for i, target in enumerate(mnist.target[60000:])]))[:, 1]\n",
    "    mnist.data[:60000] = mnist.data[reorder_train]\n",
    "    mnist.target[:60000] = mnist.target[reorder_train]\n",
    "    mnist.data[60000:] = mnist.data[reorder_test + 60000]\n",
    "    mnist.target[60000:] = mnist.target[reorder_test + 60000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(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.]]),\n",
       " array([0, 0, 0, ..., 9, 9, 9], dtype=int8))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "try:\n",
    "    from sklearn.datasets import fetch_openml\n",
    "    mnist = fetch_openml('mnist_784', version=1, cache=True)\n",
    "    mnist.target = mnist.target.astype(np.int8) # fetch_openml() returns targets as strings\n",
    "    sort_by_target(mnist) # fetch_openml() returns an unsorted dataset\n",
    "except ImportError:\n",
    "    from sklearn.datasets import fetch_mldata\n",
    "    mnist = fetch_mldata('MNIST original')\n",
    "mnist[\"data\"], mnist[\"target\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 784)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mnist.data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 784)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X, y = mnist[\"data\"], mnist[\"target\"]\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000,)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "784"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "28*28 #resolución de las imágenes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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": [
    "some_digit = X[36000]\n",
    "some_digit_image = some_digit.reshape(28, 28)\n",
    "plt.imshow(some_digit_image, cmap = mpl.cm.binary,\n",
    "           interpolation=\"nearest\")\n",
    "plt.axis(\"off\")\n",
    "\n",
    "save_fig(\"some_digit_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_digit(data): #función graficar dígito\n",
    "    image = data.reshape(28, 28)\n",
    "    plt.imshow(image, cmap = mpl.cm.binary,\n",
    "               interpolation=\"nearest\")\n",
    "    plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [],
   "source": [
    "# EXTRA\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 = mpl.cm.binary, **options)\n",
    "    plt.axis(\"off\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure more_digits_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 648x648 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "(100, 784)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "plt.figure(figsize=(9,9))\n",
    "example_images = np.r_[X[:12000:600], X[13000:30600:600], X[30600:60000:590]] #concatenar a la derecha\n",
    "plot_digits(example_images, images_per_row=10)\n",
    "save_fig(\"more_digits_plot\")\n",
    "plt.show()\n",
    "\n",
    "np.shape(example_images)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y[36000] #etiqueta correspondiente"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = X[:60000], X[60000:], y[:60000], y[60000:] #particionar train y test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "shuffle_index = np.random.permutation(60000) #permutación datos\n",
    "X_train, y_train = X_train[shuffle_index], y_train[shuffle_index]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Ejemplo clasificador binario"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_5 = (y_train == 5) # clasificador 5 vs NO 5\n",
    "y_test_5 = (y_test == 5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Nota**: Algunos parámetros de optimización deben definirse para evitar warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
       "       early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
       "       l1_ratio=0.15, learning_rate='optimal', loss='hinge', max_iter=5,\n",
       "       n_iter=None, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
       "       power_t=0.5, random_state=42, shuffle=True, tol=-inf,\n",
       "       validation_fraction=0.1, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.linear_model import SGDClassifier #clasificador lineal por gradiente\n",
    "\n",
    "sgd_clf = SGDClassifier(max_iter=5, tol=-np.infty, random_state=42)\n",
    "sgd_clf.fit(X_train, y_train_5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True])"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.predict([some_digit]) #clasificar dígito 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.87175, 0.96715, 0.9369 ])"
      ]
     },
     "execution_count": 32,
     "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\") #evaluar clasificador según acierto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.87175\n",
      "0.96715\n",
      "0.9369\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.base import clone\n",
    "\n",
    "skfolds = StratifiedKFold(n_splits=3, random_state=42) #generar partición balanceada\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": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.base import BaseEstimator #clase básica para construir modelos en scikitlearn\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": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.91025, 0.9101 , 0.9086 ])"
      ]
     },
     "execution_count": 35,
     "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": 46,
   "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) #genera predicción de cada muestra según su aparición en validación"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[50766,  3813],\n",
       "       [  671,  4750]], dtype=int64)"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "confusion_matrix(y_train_5, y_train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_perfect_predictions = y_train_5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[54579,     0],\n",
       "       [    0,  5421]], dtype=int64)"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(y_train_5, y_train_perfect_predictions)\n",
    "#[[TN FP][FN TP]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5547121335980381"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "\n",
    "precision_score(y_train_5, y_train_pred) # TP/(TP+FP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.876222099243682"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred) # TP/(TP+FN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6793478260869565"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "f1_score(y_train_5, y_train_pred) # 2/(1/pre + 1/rec) = TP/(TP + (FN+FP)/2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([75090.39935667])"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores = sgd_clf.decision_function([some_digit]) #estimar score del dígito de prueba\n",
    "y_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "threshold = 0\n",
    "y_some_digit_pred = (y_scores > threshold)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ True])"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_some_digit_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([False])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 200000 #cambiar umbral para sesgar clasificador según precisión o exahustividad\n",
    "y_some_digit_pred = (y_scores > threshold)\n",
    "y_some_digit_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "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": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000,)"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_scores.shape  #revisar salida de predict según versión de scikitlearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# hack to work around issue #9589 in Scikit-Learn 0.19.0\n",
    "if y_scores.ndim == 2:\n",
    "    y_scores = y_scores[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import precision_recall_curve\n",
    "\n",
    "precisions, recalls, thresholds = precision_recall_curve(y_train_5, y_scores) #estimar precisión y exahustividad variando los umbrales"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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.xlabel(\"Threshold\", fontsize=16)\n",
    "    plt.legend(loc=\"upper left\", fontsize=16)\n",
    "    plt.ylim([0, 1])\n",
    "\n",
    "plt.figure(figsize=(8, 4))\n",
    "plot_precision_recall_vs_threshold(precisions, recalls, thresholds)\n",
    "plt.xlim([-700000, 700000])\n",
    "save_fig(\"precision_recall_vs_threshold_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "(y_train_pred == (y_scores > 0)).all() #revisar si todos cumplen condición\n",
    "#por defecto scikitlearn evalua clasificador con umbral en 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_pred_90 = (y_scores > 70000) # nueva frontera cambiando el umbral de 0 a 70000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6681985294117647"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision_score(y_train_5, y_train_pred_90)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8046485888212507"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred_90) # por nuevo umbral se ajusta exahustividad"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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",
    "\n",
    "plt.figure(figsize=(8, 6))\n",
    "plot_precision_vs_recall(precisions, recalls)\n",
    "save_fig(\"precision_vs_recall_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Curva ROC: reciever operating characteristic\n",
    "Característica Operativa del Receptor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1129646.83463613,  775216.34718599,  769140.56646211,\n",
       "        723908.29365335,  709589.41644501])"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_curve\n",
    "#matriz confusión = #[[TN FP][FN TP]]\n",
    "fpr, tpr, thresholds = roc_curve(y_train_5, y_scores) # Fpr = FP/N -> 1 - especificidad , Tpr = TP/(TP+FN) -> exahustividad (recall, sensibilidad)\n",
    "#especificidad = TN/(TN+FP)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure roc_curve_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAGoCAYAAABL+58oAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3Xd4lFX6xvHvmUkPKfReVYqAdFEQCyvqqqiIBdSIymJB3cW1rCJrQRT1Z91FWWXtrB1YGwK6dhGUqhRFiiC9hPQymZnz+2MmMSCBCZnknUzuz3VxOfPOm5kbdkkeTnmOsdYiIiIiEk1cTgcQERERCTcVOCIiIhJ1VOCIiIhI1FGBIyIiIlFHBY6IiIhEHRU4IiIiEnVU4IiIiEjUqfECxxhzgzFmkTGm2Bjz4iHuvckYs90Yk22Med4YE19DMUVERKQWc2IEZyswCXj+YDcZY04Hbgf+ALQDOgD3Vnc4ERERqf1qvMCx1s601v4X2HOIW0cBz1lrV1pr9wL3AVdUdz4RERGp/WKcDnAQXYF3yj1fDjQ1xjS01u5THBljrgauBkhOTu7TuXPnmkspIiLiEAv4rcXns1igxOff50Vb7okNXvP4/LiMCT63FHn9uI3Z5z3LHgePc/L5LSW+314pLPHhMmCCX2etxesP79FPnu1rd1trGx/u10dygVMPyC73vPRxCvuN/lhrnwWeBejbt69dtGhRjQQUEZHoVuLz4/VZ/Lb0V+CHud9Sdi2n0EuJz4/Pb7EWfMHrRR4fS3/NIj7GhS27P/Df0vfw+S3fb86iRXoiRSV+Fm3MpE2DJEp8fhasz6R9o+RAAVP63v7Ae+/MLcYYONzjJMuVQYS6uDWx3OPUg90X66awxEeP1um4DbhdpqwQ2pVbTO829dmZW0THpikkxbkxxuA2Br+3mFlTJ9P2mOP549nncOlxbTdW8re1j0gucPLY98+w9HGuA1lERMRh1lr2FpSwJ68Yrz/wQz+/2EtWYQkuY4JFgA0WGLBuZx6pibGU+Pz8vCOP5Hg3BgKFRbDI8Pkti37ZS5uGSRSV+Fi5NYf4GBclPkt2YYkjv8+NewrKHm/YnV/hfdaCMRDndlHs9dOhUTIlfj85hV66tUzFFSwqXMbgMoH/GmMwBjbuyadvuwbEugxul4utWYV0b5WG2xW41xC4r+xrXQavz3JEk3okxLiIcQfeq3G9eNwuU/YrMdZNcnzlS4vVq1czYsSlfP/995zapxOX9G/DpZX/o9tHJBc4K4EewJvB5z2AHftPT4mIiHN8fsveAg9en6XE52dXXnHZSIPPbyks8bIrt5i4GBdeX+Damh15pCbG8MvufEp8loRYd9loSOlIxdqdefy0I5dmqQl4/X48Xj85Rd5q+32sL1dIHOhf0Ulx7mCB8NsPfRP8b6HHR77HR5fmqbhdpa8HXssr9lLis/yhcxNcrv2KhmDBgbUUlvg4skk9jDHEug1NUxKIcbtIiHWRmhAb+Jrge7uD75OWGEt8jLva/kxqgrWWF154gRtvvJHk5GRmz57NH//4x7C8d40XOMaYmODnugG3MSYB8Fpr9/9/7svAi8aY/wDbgAnAizWZVUQkklhr95nqsOz7PL/YS6HHFxjB8AdGMjxeP7vzivH5Yd2uPOrFxwTu91t8Fvx+y+78YrILStiVW4zPWuLcLvzBNRWLftlLq/qJWAtev591u/KJi3HhMuD1hX/dxf625xQd8Hqnpim4XYYYt+HXzAI6Nk0hNTEWV3BKxFVuSqRH63RiXIaducUc3Ty1bPTBXW50wuP107ZhErFuFy3SEklNjMHtMsTHuImLUcu46vLZZ58xevRoBg8ezPTp02nevHnY3tuJEZwJwN3lnl8G3GuMeR5YBRxtrd1krZ1jjHkY+JTA1N+M/b5ORMQxfr+loMQXXE/x+3UZpYXH7lwPG/YERgfW7swjOc5dVoB4/ZbcIi9zVmwnPSmWn3fm0TQ1Hr+fciMasDuv2NHf64/b9x3T8Hj9v7unSUo8sW4X2YUlJMS6aNcwuWzaIjPfQ/2kOFqkJxLjMrhchp05RfRqk05usZdW9ZNIinWXjUyUFig+v+XIJvVoWC+OWJcLt9uQEh9Ttp5Daq/s7GzS0tI4+eSTmTVrFkOHDsXtDu9olCldIR0ttMhYRMorLTwKPF7yir1lC0a9fsuu3GL8fkuJ37JhVx7xsW5+3JZDvYQYtmUV8clPO2nX8LdFnj6/ZXdeMbvzPI79fkqnOAz8bsqk2OvH67e0b5RcNgXiDo5ctK6fSP3kOPKKvHRunoI7OD1SWoTkFnlpnBJPQqyL1vWTSAwWHG5XoJhokhpPTHC9RlKcm9SE2MAISrBgEQmF3+/nscce4/7772f+/Pl06dKlwnuNMYuttX0P97MieQ2OiNRx1lrW785n3c489hZ42J3n4dfMAuJiAlMohR4/y37dS+sGSazblUehx0dKQmzZwswYl6nyFMqygqyDvl4vPuaA6yoCazSgxGfJKvDQpXkqnZqmsDmrkF5t0nGXKy4A3MZwbPsGNElNKCsayk+hJMa6y3aciNRGO3fuZNSoUcyZM4fzzz+fZs2aVevnqcARkQpZa/H4Ags8d+UWk1ccWDBaOn1QuhMlp8hLZr6HxFh32TbW0umVFVuyaZqaQInPz6KNe2mZnkBesZclG7No3ygZr99ftl12U2ZB2Zba0uuhWLfrtwWi5UdXSoub0gWhPr+lZXpiWWFRXOKjyOuna4tUjDHszffQq006u3KL6dk6HZcxNEiO44gm9QLFhiswfeI2hubpidQ7jN0iInXRxx9/TEZGBnv37mXq1Klcc8011V6s62+nSC2UU1RCiTfQd6PY6ycz34MxpT0yKNv+6rfw3S+ZbMsuotDjDexk8QfWf6zelktyvJvYcgtKV2/LISHWTUKMm+zCEgpLfGHPvnpbTtnjn3b8fr9K8QHWdwD0bVufhFg3LdITsBa6tkglLsaNMRDrdtEsNQFjoFlaAnFuFwmxbtISYzWFIhIB3n33XerXr8+8efPo3r17jXymChyRGpSZ72FbdiF+f6APR+m6jm837KHA48PrD+x6WbwxsHPl+83ZJMS6WLcrn5SEGLw+Wy1FR3lFJX5g3/4fSXFuCjw+GtWLx28treon0qhefHCNR2BaZm+BhwbJcTRJSSibrilt8JVTVEKnpinEuA3FJX7aN07GZQxpibEkx7lxla7lMIaUhBgS49zEuFz7dEoVkdrll19+ITs7mx49evDwww/j8/lITk6usc9XgSMSJvnFXtbuzGP1thx81vJrZiHZhR5e+/ZXUhNiKt3D44ct2fs8z93v65PiAg21YlyGbdlFHNE4mXrxMbhcv/XgKO2p8cvuAs7o1ozW9RMDX+MO3FNc4qdV/URi3C7cLnC7XLiNoWG9OGLchni3m7Sk2Cr/2YhI3TJjxgxGjx5Nhw4dWLx4MQkJCTWeQQWOyCHkF3vJLQq0Yv9m/R5yCktYuimLbdmFLNmURYPkODLzD76rZv/ipmuL1LKtsKXdP2Pdhr7tGhAf48LtMvgttGmQhAHaNEyiYXIcCbFuYtyGOLdLIxsiEnEKCwu56aabeOaZZzj22GN57bXXHPtepQJH6ixrf1u/kl/sZXtOET9tz6XQ42P2iu3szitmV+6h+4+UL25SEmJolpqA22U4/oiGNE9LoH5SHCkJsQw4siFJsW5i3GoaJiLRZ9u2bZx22mmsWLGC2267jfvuu4+4uDjH8qjAkahS7PWxLauINTtyWbUthx+35ZKeFFt2bs0PW7JZuzOv7DC4ymiZnkhRiY89+R6uPekIir0+ujRL5fgjGpKWFEtyXEzZll8RkbqmSZMmdOnShUceeYTTTz/d6TgqcCTy5RV7KfB4Wbsjj6zCEjbszsdlDGt35rHs171s3FNAs7QENu8tDPk9yxc3cW4XHp+fLs1T2ZVbRIdG9ejXvj5HNUnhyCb1aNcoWduBRUQOIDs7m9tvv527776bZs2a8eabbx76i2qIvmtLRCn0+Jj9wzY+X7OLBev3sCffgy+EZij7FzdHNE6mxBdo835Uk3oc0aResAurwVro3DyFZqkJpCc5N3wqIlKbffvtt4wYMYJNmzZxyimncNFFFzkdaR8qcMQxBZ7ArqMZizezdlceP+/IY2cFa14aJsexJ99DXIyLwZ2akFng4dh2DfBZS8em9TiqSQqN6sWTnhRLQmztPl1XRCSS+f1+Hn30UcaPH0/Lli358ssvOf74452O9TsqcKRGWGtZ+msW3/+axX+XbWXZrwdvf9+hcTInd2zCse0bcHKnxipaREQixOTJk5kwYQIXXHAB06ZNIz093elIB6QCR8LC6/Ozelsu7yzbwoqt2fy4PZfUhFg2ZRaUNYmrSMem9ejaIo3TuzalW8s0WqYnagu0iEiEKSkpITY2luuuu46WLVsyatSoiP5erQJHQrIzt4gfNmezYksOMW7D12t3Bxra+W2FDeyyCgLdcMsXN/XiY7igTysS49yM6Neatg1rrquliIhUXklJCXfddRefffYZX3zxBQ0aNOCKK65wOtYhqcCRA8ouKOHj1Tt4cf4v/LIn/3dddCviMnBMq3TO69mCjk1TaJqWQNPUBOJjXMQE2/aLiEjtsGHDBi655BIWLFjA1VdfjdfrJTa2dnQ3V4EjAKzcms036/awK7eYtxZvrrAzb8v0RI5sUo8erdPx+y3dW6XRrWVaWZddERGJDm+99RZjxozBWsubb77JhRde6HSkSlGBU0flFXv5eu1u/rNwE1+v3V3hVuzLjmtDy/QkTu/alA6N69VwShERcYLH4+Guu+6ic+fOvPbaa7Rv397pSJWmAqcOeeLjNfy4LZc5K7dXeM8fOjehe6s0erepz4kdG9dgOhERcdrq1atp06YNycnJzJs3j2bNmtWaKan9qcCJcvnFXq77zxK+WLPrgK8nxrq5qG8rbhrSUU3vRETqKGst06ZNY9y4cVxzzTU8/vjjtG7d2ulYVaICJwp5vH6WbtrLHTN/YP3u/H1eO6pJPe44szPtGiZryklERMjKyuLqq6/mrbfeYsiQIfztb39zOlJYqMCJAtZa/rNwE4s37mXB+j1syy763T1nH9Ochy84hqQ4/U8uIiIBS5cu5fzzz2fz5s089NBD3HLLLbhcLqdjhYV+2tViRSU+7nt/Ff9ZuOmAr7dukMhFfVpz9UkdiI/RDicREdlXWloa6enpvP766/Tv39/pOGGlAqcWWrcrjwumzmdvsJFeqcYp8Uw6rxu9WqfTJDXBoXQiIhLJtm/fznPPPcf48ePp0KEDS5YsicoeZSpwapGfd+Ry3werf7dg+IoB7bjtjE6afhIRkYOaN28eGRkZ5ObmMmzYMI4++uioLG5ABU7EW/RLJlM+XctnP/1+F9Rf/nAUNw3p6EAqERGpTUpKSpgwYQIPP/wwXbt25ZNPPuHoo492Ola1UoEToTbtKeDE//v0gK8N7tyEqZf11roaEREJyQUXXMC7777Ltddey2OPPUZiYqLTkaqdCpwIk11YwjWvLGLB+syya26XYegxzbn5tE60bpDkYDoREalNrLUYY7jxxhvJyMjgggsucDpSjVGBE0FeXbiJ8bN+2OfaW9ceT792DRxKJCIitVFBQQHjxo2jZcuW3H333Zx66qlOR6pxKnAiQG5RCd3vmbfPtTGD2nPbGZ2JdUdHPwIREakZK1as4OKLL2b16tWMHz/e6TiOUYHjsP+t3sHolxaVPe/ZOp0XruhH/WQdmyAiIqGz1vLss88ybtw40tLSmDt3LkOGDHE6lmNU4Djoyhe+5dNyu6PuPLMLY07s4GAiERGprX7++WduuOEG/vCHP/DSSy/RtGlTpyM5SgWOA77dkMlVL35HXrEXgPpJsSwY/wftihIRkUrbuHEjbdu2pWPHjsyfP58+ffpEzXELVaE/gRp27SuLueiZb8qKm1O7NGHRhCEqbkREpFL8fj+TJ0/myCOPZPbs2QD069dPxU2QRnBqSInPz8n/9xlbsgrLrj0wrDsjj20dtV0kRUSkemzbto2MjAz+97//cfHFFzNw4ECnI0UcFTg1YGduEcfe/7+y56d2acqzGX1wuVTYiIhI5cydO5eMjAzy8vL497//zVVXXaV/KB+ACpxq9ulPO7nyhe/Knl9/yhHcenpnBxOJiEhttnXrVpo1a8brr78e9cctVIWx1jqdIaz69u1rFy1adOgba8CvmQUMevi34xb+fXlfTj26bq9qFxGRylu/fj0rV65k6NChWGspKSkhLi6624kYYxZba/se7tdrJVI1WbxxL2f+48uy5wvu+IOKGxERqbTXX3+dnj17MnbsWIqLizHGRH1xEw4qcKrBJdMWMHzqfHKLAjulZo4dQLO0BIdTiYhIbZKfn8/o0aMZOXIk3bt356uvviI+Pt7pWLWG1uCE2Qffb2P+uj1lz7+49RTaNNQBmSIiErq8vDyOPfZYfvzxR+68807uueceYmL0I7sy9KcVRpn5Hm59ezkAvdqkM/O6AVrZLiIilVavXj0uvvhiBg0axODBg52OUytpiipMvD4/x03+HwUeH/ExLqaP7q/iRkREQrZ3714uvvhivv32WwDuvvtuFTdVoAInTB6a8yMerx+A9248geR4DY6JiEhovv76a3r27MnMmTNZsWKF03GiggqcMFj2axbTvtwAwH3ndaNj0xSHE4mISG3g8/m4//77Oemkk4iJiWH+/PlcddVVTseKCipwwuCOmT8A0LtNOhnHtXU4jYiI1BYvv/wyEyZM4KKLLmLp0qX069fP6UhRQ/MoVZSZ72HdzjwAbj6tk8NpRESkNsjKyiI9PZ3LL7+chg0bMnToUK3bDDON4FTRgx+uxuPz0799AwYe2cjpOCIiEsE8Hg8333wzXbp0YceOHbjdbs455xwVN9VAIzhV8MZ3m3hz0WYArhzYztkwIiIS0dauXcvIkSNZtGgR119/PWlpaU5HimoqcA7Tii3Z/G1GYO1No3rxnNGtucOJREQkUr366qtcc801xMbGMnPmTIYNG+Z0pKinAucwnf3Pr8oev3fjQAeTiIhIJLPW8uabb9KjRw9effVV2rRp43SkOkEFzmG46Y1lZY/fuvZ4mqclOphGREQi0fLly0lNTaV9+/a88sorJCYm6riFGqRFxpX0845cZi3dAsAVA9rRr10DhxOJiEgksdby1FNP0b9/f2666SYAUlJSVNzUMP1pV9K4cqM395zT1cEkIiISaTIzMxk9ejT//e9/OfPMM5k2bZrTkeosFTiV8MPmbFZuzQHgpauOdTiNiIhEklWrVnHGGWewfft2HnvsMf7yl7/gcmmixCkqcCrh6c/WAnB+75ac1LGxw2lERCSStG3blmOOOYaZM2fSt29fp+PUeSotQ7RxTz4frtgOwKjj2zkbRkREIsKWLVsYM2YM+fn5JCcn8/7776u4iRAqcEI0efaPAHRulkKP1ukOpxEREad98MEHZVu/ly5d6nQc2Y8KnBD8uD2HOSsDozf3ndfN4TQiIuKk4uJibrrpJs4++2xatWrFkiVLOOGEE5yOJftRgROCf/zvZwBapCVoW7iISB13/fXX88QTT3DjjTeyYMECOnXSQcuRSIuMD2H1thxm/xAYvZk8/BiH04iIiFM8Hg9xcXGMHz+eoUOHcu655zodSQ5CBc4hPP7RmrLHA49o6GASERFxQl5eHjfccANZWVnMmjWLDh060KFDB6djySFoiuog/H7LvFU7AHgmow8xbv1xiYjUJcuWLaNPnz68/PLLHHPMMfj9fqcjSYj0E/sgFm/aC4DbZTi9azOH04iISE2x1vLPf/6T/v37k5eXxyeffMLEiRNxu91OR5MQ1XiBY4xpYIyZZYzJN8ZsNMZcUsF98caYfxljdhhjMo0x7xljWtZk1tIzp/q0qV+THysiIg7bu3cvkyZNYsiQISxfvpyTTz7Z6UhSSU6M4DwFeICmwKXAVGPMgQ51+gtwPHAM0ALIAv5ZUyEBFqzfA8BF/VrX5MeKiIhDFi9ejM/no0GDBixcuJD33nuPRo0aOR1LDkONFjjGmGRgOPB3a22etfYr4F0g4wC3twfmWmt3WGuLgNeBGjvd8uu1u1m/K5+4GBdDezSvqY8VEREH+Hw+7r33Xo499lj++c/Av6XbtWuHMcbhZHK4anoXVUfAZ61dU+7acuCkA9z7HPCkMaZ09OZS4MPqjxgw+cPVAAzv3ZL4GM25iohEq82bN3PZZZfx+eefk5GRwejRo52OJGFQ0wVOPSB7v2vZQMoB7l0DbAK2AD7gB+CGA72pMeZq4GqANm3aVDmkx+tnxZbAqeGX69wpEZGoNW/ePEaOHElxcTEvvfQSl19+udORJExqeg1OHpC637VUIPcA904FEoCGQDIwkwpGcKy1z1pr+1pr+zZuXPVTvpdvzip73KX5/nFFRCRapKWlceSRR7JkyRIVN1GmpgucNUCMMeaoctd6ACsPcG8P4EVrbaa1tpjAAuNjjTHVvtrrpfm/ADDyWC0uFhGJNmvWrOGJJ54AoH///ixYsICOHTs6nErCrUYLHGttPoGRmInGmGRjzEDgXOCVA9z+HXC5MSbNGBMLjAW2Wmt3V2fGAo+X97/fBsBlx7Wtzo8SEZEa9vLLL9O7d28mTZrE7t2BHydaSBydnNgmPhZIBHYCrwHXWWtXGmMGGWPyyt13C1AE/AzsAs4EhlV3uOkLNgKQkhBD1xZp1f1xIiJSA3Jzc8nIyGDUqFH06dOHZcuWaft3lKvxs6istZnAeQe4/iWBRcilz/cQ2DlVoz5evROAPw8+6hB3iohIbeDz+TjxxBP5/vvvueeee5gwYYI6EtcBOmyznB05RXy7IROAs9X7RkSkVrPWAuB2u/nb3/5GixYtOPHEEx1OJTVFZ1GVc+esFQCkxMfQPC3R4TQiInK4du/ezTnnnMNLL70EwIgRI1Tc1DEqcMr5eHXg5PDb/tjZ4SQiInK4Pv/8c3r06MG8efMoKipyOo44RAVOUHZhSdnjC/u0cjCJiIgcDq/Xy913383gwYNJTk5mwYIFXHvttU7HEoeowAn6Lrj2JiHWRUKsFp+JiNQ2X375JRMnTiQjI4MlS5bQq1cvpyOJg7TIOOjO//4AwOgT2jucREREKmPDhg20b9+eU045hYULF3Lsscc6HUkigEZwCDT325FTDMBFfdW9WESkNigqKuLGG2+kU6dOLFu2DEDFjZTRCA7w2U+7yh63bZjsYBIREQnFTz/9xMUXX8zy5csZN24cXbp0cTqSRBgVOMB7y7cCcFZ39b4REYl0L7/8MmPHjiUhIYH333+fs846y+lIEoFU4AAfrtgOwKXHtXE4iYiIHMq6devo168f06dPp2XLlk7HkQhV5wuc7zdnlT3u3aa+g0lERKQiixcvJi8vj5NOOom77roLQMctyEHV+UXGz36xHoC2DZO0PVxEJMJYa3n88cc5/vjjueWWW7DW4na7VdzIIdX5Aqd0euqvQzo6nERERMrbtWsXQ4cO5a9//Stnnnkmc+bMwRjjdCypJer0FFVWgafs8cmdmjiYREREytu8eTP9+/dn9+7dTJkyhbFjx6q4kUqp0wXOyq05+PyW1g0SSUuMdTqOiIgEtWzZkosuuohRo0bRs2dPp+NILVSnp6h+2p4LQL+2DRxOIiIimzZt4qyzzmLDhg0YY3j88cdV3Mhhq9MFzvx1ewBo10jN/UREnDRr1ix69uzJF198wU8//eR0HIkCdbrA2ZJVCECr+okOJxERqZuKioq4/vrrOf/88+nQoQNLly7ljDPOcDqWRIE6W+D4/JaNe/IB6KX+NyIijpg8eTJPP/00N998M/Pnz+fII490OpJEiTq7yHjJpr0UeHy0SEugvaaoRERqjLWWrKws6tevz2233cYJJ5zAkCFDnI4lUabOjuB8vHoHAMe0Snc4iYhI3ZGTk8Oll17KwIEDKSgoIDk5WcWNVIs6W+AUl/gBaJGu9TciIjVh0aJF9O7dmzfeeINLLrmE+Ph4pyNJFKuzBc6aHYEt4r3bagRHRKQ6+f1+Hn30UQYMGIDH4+Hzzz9nwoQJOm5BqlWdLHCstSzeuBeAHpqiEhGpVj6fjzfeeIOzzz6bZcuWccIJJzgdSeqAOrnIePW2XIq9fhokx9G6QZLTcUREotJnn33GMcccQ4MGDZg3bx5paWk6bkFqTJ0cwdmwO7A9PNatv2giIuHm9Xq58847GTx4MBMnTgQgPT1dxY3UqDo5gvP5mp0AnNm9ucNJRESiy8aNG7nkkkuYP38+V111Fffff7/TkaSOqpMFTukITpdmqQ4nERGJHl988QXnnnsuPp+PV199lZEjRzodSeqwOjdFZa3lhy3ZAAzq2MjhNCIi0aNjx44MHDiQpUuXqrgRx9W5AmfdrjyKSvykJMTQPE09cEREqmLVqlVcd911+Hw+mjVrxvvvv88RRxzhdCyRulfgrNoW6H/TNDXB4SQiIrWXtZZ///vf9O3blxkzZrBu3TqnI4nso84VOJ/+GFhgPOCIhg4nERGpnbKzsxk5ciRjxoxhwIABLF++nI4dOzodS2Qfda7AyS/2AtAsTSM4IiKHY/jw4bz99ts88MADzJ07l+bNtSNVIk+ldlEZYzoBg4CGwIvW2h3GmNbAHmttQXUEDLfNewsBOLZdA4eTiIjUHn6/H5/PR2xsLJMnT6akpIQBAwY4HUukQiEVOMaYWOB54BLAABb4CNgBTAFWAuOrKWNYbc0OFDht1MFYRCQkO3bsYNSoUXTq1Iknn3ySfv36OR1J5JBCnaK6DzgHGAO0JVDklJoNnB7mXNXC4/WTVVCC22VoVE+n2IqIHMrHH39Mjx49+Oyzz+jSpYvTcURCFmqBcynwd2vt88DW/V5bD7QPa6pqsj27CID6SXG4XGoZLiJSkZKSEsaPH89pp51GgwYN+O6777j22mudjiUSslALnMbAioO8XitW7C7YsAeAVvXV/0ZE5GA2btzIk08+yZ/+9CcWLVpE9+7dnY4kUimhLjLeCPQDPjnAa32Bn8OWqBpl5nuA33ZSiYjIvhYsWED//v0eONspAAAgAElEQVQ58sgjWbVqFW3btnU6kshhCXUEZzpwpzFmOOAOXrPGmOOBvwIvVkO2sNuRE5ii6t9BO6hERMorLCzk2muv5fjjj+edd94BUHEjtVqoIziTgd7AW0Be8NqnQAowC3gi/NHCb+OewE723m3qO5xERCRyrFy5kosvvpiVK1dy2223cdZZZzkdSaTKQipwrLVeYJgxZgiBHVNNgD3AHGvt3GrMF1afBLsYH9G4nsNJREQiwyuvvMI111xDSkoKc+bM4fTTa8WmWJFDCrUPThMCzfw+ItD/pvxrLqCRtXZnNeQLG2stqQkx5BR5dQ6ViEhQSkoKJ5xwAi+//DLNmjVzOo5I2IS6Bmcb0KeC13oFX49oW7IKySnyEh/jommqeuCISN21YMECnn/+eQDOO+885s6dq+JGok6oBc7BmsbEAP4wZKlWP2zOBqBL81SMUQ8cEal7/H4/Dz30EIMGDeLBBx+kuLgYQN8TJSpVOEVljKkHpJa71MgY02K/2xIJHN+woxqyhdWGPfkAFHsjvhYTEQm7HTt2kJGRwUcffcQFF1zAtGnTiI/XaLZEr4OtwbkZuCv42ALvVXCfAe4PZ6jq1FpN/kSkjsnNzaV3795kZmbyzDPPMGbMGI3aSNQ7WIHzPrCdQAHzNPAwsGG/e4qBVdbab6snXvj8mhk8Rby9euCISN1grcUYQ0pKCnfddRcDBw6kW7duTscSqREVFjjW2sXAYgBjjAVmWGt311SwcNuwO9C+54gm2iIuItFvw4YNXHrppUycOJFTTz2Va665xulIIjUqpEXG1tpnanNxA78d09A0RVvERSS6vfnmm/Ts2ZNVq1aRn5/vdBwRR4TayRhjTEfgSqATvz9c01prI7r15bbgSeIpCSH/lkVEapWCggLGjRvHtGnT6N+/P6+99hrt27d3OpaII0Jt9NcH+JLAbqk2wE9AAwIdjbcCm6orYLjkFgUO2EyKcx/iThGR2unNN9/k3//+N7fffjsTJ04kNjbW6Ugijgm1D86DwAfAUQQWHV9mrW0GnB18j79VT7zwKPb6yh7XT4pzMImISHhZa1m/fj0Ao0aN4rvvvmPy5MkqbqTOC7XA6UHgxPDSJjJuAGvtbOABAjusItbuPE/ZY5dLWyNFJDpkZWVx4YUX0qtXLzZv3owxhj59Kmo6L1K3hLogJR7Itdb6jTGZQNNyr60Cjgl7sjBasyMXgKO0g0pEosQ333zDyJEj2bJlC/fffz8tWuzfh1Wkbgt1BGc9UPq3ZyVwRbnXLgMi+qDNjbsDuwgKPL5D3CkiEtmstTz44IMMGjQIYwxfffUVt912Gy5XqN/OReqGUP9GfAgMCT6eDJxrjMk0xuwERgH/qI5w4fLr3kCTv+G9WzqcRESkaowxrFmzhuHDh7N06VL69+/vdCSRiBTSFJW1dny5x3OMMYOAC4AkYI619t1qyhcWWQUlALRI1zENIlI7zZ07lxYtWtC9e3eeeeYZYmJidNyCyEEc1pimtXaBtfYWa+3YSC9uAHbkBHrgNE1Tkz8RqV08Hg+33XYbZ5xxBhMnTgQgNjZWxY3IIVR50tYYc7Qx5rVwhKkuu3KLAWiSopNzRaT2WL9+PYMGDeL//u//uPbaa3n55ZedjiRSaxx0isoE/onQnUBzv3XW2tXlXutO4LTxYUBhdYasqt15gQKnQbJ64IhI7bBkyRJOOeUUXC4Xb7/9NsOHD3c6kkitUuEIjjGmGfA1sBR4B1hhjHnJGBNjjJkSvH42gZPGj6yJsIfDWsue4DlUKnBEpLbo1q0bl1xyCUuXLlVxI3IYDjZF9SDQE7gfGA7cApwGfA6MBd4CjrLW/tlau6O6gx6unOARDQDxMTqmQUQi1w8//MAZZ5xBZmYmcXFxTJ06lXbt2jkdS6RWOtgU1RBgorX2wdILxpgVwFzgX9basdUdLhx2BhcYt9QOKhGJUNZa/vWvf3HTTTdRv359fvnlFxo0aOB0LJFa7WAjOE0ITFGVV/r8sBcVG2MaGGNmGWPyjTEbjTGXHOTe3saYL4wxecaYHcaYv1T283bkBNbftEjXDioRiTx79+7lggsuYOzYsZxyyiksX76c3r17Ox1LpNY7WIHjBor3u1b6PL8Kn/kU4CFw3MOlwFRjTNf9bzLGNALmAM8ADQms85lX2Q/LLAisv2mYrB1UIhJ5xo0bx7vvvssjjzzCBx98QJMmTZyOJBIVDtXo7zRjTPkFxC7AAmcYYzqXv9Fa++qhPswYk0xgPU83a20e8JUx5l0gA7h9v9v/Csy11v4n+LwYWE0l5RcH1uCkJoZ67JaISPXy+Xzk5uaSnp7OQw89xA033EC/fv2cjiUSVQ71U39iBdcn7ffcAocscICOgM9au6bcteXASQe49zjgB2PMfAKjNwuB6621m/a/0RhzNXA1QJs2bfZ5bUPwHKrUhNgQ4omIVK9t27aRkZGBz+fj448/plmzZjRr1szpWCJR52AFTpdq+Lx6QPZ+17KBlAPc2wroTWCx8w/AwwTW/gzc/0Zr7bPAswB9+/a15V/bmhVo0ZOaqAJHRJz14YcfMmrUKPLy8pgyZYoOyBSpRhUWONban6rh8/KA1P2upQK5B7i3EJhlrf0OwBhzL7DbGJNmrd2/SKpQYfAE8Ub1tAZHRJzh8XgYP348jz76KN27d+eNN96gS5fq+DekiJSq6X8+rAFijDFHlbvWA1h5gHu/JzD1Var0caUOYNmRG9gm3rFpvcp8mYhI2BQWFjJz5kzGjh3LwoULVdyI1IAaLXCstfnATGCiMSbZGDMQOBd45QC3vwAMM8b0NMbEAn8HvrLWZlXmM/fkBXZRNdY5VCJSw9577z2Ki4tJS0tj6dKlPPXUUyQmqieXSE1wYgJ4LJAI7CSwpuY6a+1KY8wgY0xe6U3W2k+A8cAHwXuPBCrsmXMg1lq2ZQdGcFTgiEhNyc/PZ/To0ZxzzjlMnToVgLS0NIdTidQtNb532lqbCZx3gOtfEliEXP7aVGDq4X5WXvFvxzQkxuqYBhGpfsuXL2fEiBH89NNPTJgwgRtuuMHpSCJ1UlQ3hyktcOJjXAQORhcRqT5vvPEGo0aNokGDBnz88ccMHjzY6UgidValp6iMMUcaY/obY5KqI1A4la6/SdDojYjUgO7du3P22WezfPlyFTciDgu5wDHGjDbGbAZ+AuYDnYPX3zbGXFtN+aoku7AEgIb14hxOIiLR6quvvuLWW2/FWsvRRx/N22+/TePGjZ2OJVLnhVTgGGOuINBI7xNgFPtu1V4IXBz2ZGGwN3gOVaemB+ojKCJy+Hw+H5MmTeKkk05i1qxZZGZmOh1JRMoJdQTnVuBJa+3l/P4k8dUER3MizY/bAv0DUxKieqmRiNSwrVu3MmTIEP7+978zYsQIlixZQsOGDZ2OJSLlhPqT/wgC27UPJBeoH5444ZVVGBjBKb+bSkSkKnw+H6eccgqbN2/mhRdeYNSoUdrEIBKBQi1wMoHWFbzWEdgWnjjh5fH6AejQSF2MRaRqPB4PMTExuN1unnrqKVq1akXnzhE5eC0ihD5F9QEwwRhTvsixxph0YBzwTtiThUFxsMBpmqomfyJy+NauXcuAAQN44oknADj11FNV3IhEuFALnDuD964C3idwLtQjweexwL3Vkq6KVmwJnMmpk8RF5HD95z//oVevXqxfv54jjjjC6TgiEqKQChxr7U6gN/APoDGwBWgAvAT0t9burbaEVeDxBUZw1AdHRCorLy+PK6+8kssuu4yePXuyfPlyzj33XKdjiUiIQt5eFDzk8s7gr1qlYbL64IhI5Sxbtozp06dz11138fe//52YGO3GFKlNQvoba4x5AHjZWvtjNecJK3dwZ0MDFTgiEgJrLQsXLuS4447jhBNOYN26dbRp08bpWCJyGEJdg3MjsNIYs8gYc6Mxpla06cwpCmwPT4rTv7xE5OAyMzMZNmwYAwYMYNGiRQAqbkRqsVALnCbA5cAu4DFgizHmfWPMhcaYiNyi5PX5ycwP9MFppKMaROQgvvzyS3r27Mns2bN57LHH6NOnj9ORRKSKQl1kXGit/Y+19o9AK+B2oDnwBrDDGDOtGjMelr0FJWWPY9yVPlNUROqIhx56iJNPPpn4+Hi++eYbxo0bp8Z9IlGg0j/5rbU7rLWPWWv7AH8g0Mn4qrAnq6J8dS8WkRAkJSVxySWXsGTJEo3ciESRSi9OCU5JnQdcBpxG4ODNio5xcMye/GIAjmqiLsYisq/3338fr9fLeeedxw033KARG5EoFPIIjjHmZGPMc8AOAgduNgVuAVpYa8+ppnyHLb/YB8Ce4DocEZHi4mJuuukmhg4dypNPPom1VsWNSJQKdZv4JqAl8CvwFIEt4z9VZ7Cq2p5TBECX5ikOJxGRSPDzzz+Xnfz95z//mYcffljFjUgUC3WK6iMCRc3n1RkmnEqCXYytdTiIiDhu48aN9O7dm7i4ON555x3OOSfiBp1FJMxCKnCstaOrO0i4bd5bCEDbhkkOJxERp5ROQbVt25a7776bESNG0KpVK6djiUgNqLDAMcYcC6yw1hYEHx+UtfbbsCaroqTg+VOlJ4qLSN2ydOlSrrzySqZPn063bt245ZZbnI4kIjXoYCM4C4DjgG+Djyua7DHB1yLqRMtf9hQAcERj7aISqUustUyZMoVbbrmFxo0bk5ub63QkEXHAwQqcPwKrg4/PpOICJyKV9vZTPxyRumPPnj1cddVVvPvuuwwdOpTnn3+eRo0aOR1LRBxQYYFjrZ1b7vGcmokTPq7g7ogmKRF5koSIVIMpU6YwZ84cnnzySW688UbtkhKpw0Lqg2OMWWWM6V7Ba0cbY1aFN1bVFZUE+uCkJMQ6nEREqpPP5+OXX34B4I477mDx4sX8+c9/VnEjUseF2uivM5BYwWtJQKfwxAmf5ZuzAYiP1TlUItFq8+bNDB48mJNOOom8vDzi4uLo1q2b07FEJAJU5qd/RWtwjgGyw5AlrBoHp6ZiXPpXnEg0evfdd+nRoweLFy/mvvvuo149bSgQkd8cbJv4jcCNwacWeNsYU7zfbYlAC+Dt6ol3+HblBqK2b6RveiLRxOPxcOutt/KPf/yDXr168frrr9OxY0enY4lIhDnYLqqtwOLg4yOBn4A9+91TDKwCpoY/WtVs2J0PQP0krcERiSYxMTGsXr2acePG8eCDDxIfr40EIvJ7B9tFNQOYAZQu1rvTWru+hnJVWZzbhcfn1yJjkSgxffp0Bg8eTIsWLfjggw+IjdXfbRGpWEhrcKy1I2tTcePx+vH4/MS4DAlaZCxSq+Xm5pKRkUFGRgZPPPEEgIobETmkg63BuY3AAZvbg48Pxlpr/y+80Q5fdmEJACkJMdoqKlKLLV68mBEjRrB+/Xruvfde7rzzTqcjiUgtcbA1OA8CnwHbg48PxgIRU+AUegI9cJLiQj0sXUQizezZsznvvPNo2rQpn376KSeeeKLTkUSkFjlYBZBorS3dNVVRD5yItCc/EDsxLqKOxxKRShgwYACjR49m0qRJNGzY0Ok4IlLLVLhApVxxg7W2+FC/aiZuaPKC509lFXgcTiIilfHZZ58xdOhQiouLSU9PZ+rUqSpuROSwhHpUQwdjTM9yz+ONMXcbY94yxvyp+uIdntI1OP3aNXA4iYiEwuv1cvfddzN48GDWrFnDtm3bnI4kIrVcqItUnibQ72ZZ8Pl9wE3AGmCYMcZtrX2mGvIdluISPwAJsZqiEol0v/76K5deeilffvklo0aNYsqUKepKLCJVFuoe6p7AFwAmsC3pCmC8tbYrgQXI11ZLusPk8QUKnPgYbREXiXQZGRksXbqUV155hRdffFHFjYiERagjOOnA7uDjnkBD4M3g84+AP4c5V5XkFQXW4GgERyQyFRUV4fV6qVevHs888wwul4ujjjrK6VgiEkVCHeLYCXQIPh4CbLDWbgw+TwZ84Q5WFfmeQIGTkqBt4iKR5scff+S4447juuuuA6BTp04qbkQk7EItcN4H7jfGTAJuYd/DNbsCG8IdrCqKvVqDIxJprLW8+OKL9OnThy1btjBixAinI4lIFAt1iON2IAW4GPgYmFTutYuAT8Kcq0qKSgIDSlqDIxIZcnJyuO6663j11Vc55ZRTmD59Oi1atHA6lohEsZAKHGttDpBRwWv9wpooDNTJWCSyZGVlMW/ePO677z7uuOMO3G6NropI9apUBWCMSQGOBRoAe4DvrLW51RGsKn6botIIjohT/H4/M2fO5Pzzz6dNmzasW7eO1NRUp2OJSB0RcgVgjJkAbAPmAW8QmKraZoyJuNPvVm3NASA+Rv9KFHHCrl27GDp0KBdeeCHvvPMOgIobEalRIY3gGGOuByYC/wGmEziAsxlwGTDRGJNprZ1abSkrqX5yLAAeX0Rt7hKpEz799FMuvfRSMjMzmTJlCuedd57TkUSkDgp1iuoG4Glr7Q3lri0H5hpjsoEbgYgpcKwN/LdpSoKzQUTqmMcff5ybb76Zjh078uGHH9KjRw+nI4lIHRXqFFUH4J0KXnuH33rkRISyTsZagyNSo3r16sWVV17J4sWLVdyIiKNCrQAygU4VvNYp+HrEWLopC4A47dQQqXazZs3iwQcfBODkk0/mueeeIzk52eFUIlLXhVrg/JdAo78Lg2dRAWCMGUbg4M3/Vke4w9W+UeCba4zbHOJOETlcRUVFXH/99Zx//vnMmjULj8fjdCQRkTKhFji3Az8S2D1VYIzZaIwpINDR+Kfg6xGjtA9OWmKsw0lEotPq1avp378/Tz/9NDfffDNffvklcXFxTscSESkTaqO/bGPMAGAYMIhAH5xM4HPgHWttRG1X2p5TBOioBpHqkJOTw8CBA3G73cyePZs//vGPTkcSEfmdkBv9BYuYt9n3HKqIFBfjwuP1k6gCRyRsioqKSEhIIDU1leeee47+/fvruAURiVgHnaIyxowwxiwwxuw2xqw1xtxvjIn48w88wU7GOotKJDy+++47unbtyltvvQXAsGHDVNyISESrsAIwxlwIvEqgod/XQAGBtTaTKvqaSFDaAyfWbXC5tMhYpCr8fj+PPvooAwYMwOv1qqgRkVrjYEMcfwU+AI6y1p5rrT0GeAi40RgTsUMj/mCFo2MaRKpm586dnHXWWdxyyy0MHTqUZcuWMXDgQKdjiYiE5GCFSidgqrW2pNy1fwCJQNtqTVUFpQVOSkLEz6SJRLRPPvmETz/9lKeffpoZM2ZQv359pyOJiITsYAVOOrB7v2u7gv+N2O90Xn+gwEmK0wiOSGWVlJSwcOFCAEaMGMHPP//MddddR7n2VyIitcKhpppsJa87rnQNzi97CpwNIlLLbNy4kZNOOomTTz6ZLVu2ANC6dWuHU4mIHJ5DzeN8XcG/3Bbud91aa+PDlqoKbLDC6dMmYgeZRCLOjBkz+NOf/oTP5+OFF16gZcuWTkcSEamSgxU4D9VYijAqHcHRQZsih2at5YYbbuDpp5+mX79+vPbaaxxxxBFOxxIRqbIKCxxr7R01GSRcShcZx7lV4IgcijGGpKQkbr31ViZNmqTjFkQkatT4ViNjTAPgOeA0AouY77DWvnqQ++OA74F61tpWh3r/0sVBCVpkLHJA1lqee+45jj76aAYMGMDDDz+sRcQiEnWcGOZ4CvAATYFLganGmK4Huf9WYGeob166BidGTf5Efic7O5sRI0YwZswYpk2bBqDiRkSiUo0WOMaYZGA48HdrbZ619ivgXSCjgvvbA5cBk0P9jNIRnBiXpqhEylu4cCG9evVixowZPPDAAzz33HNORxIRqTY1PUXVEfBZa9eUu7YcOKmC+/8JjAcKD/amxpirgasBGrdqTxKBoxpEJGDBggUMGjSIli1b8sUXXzBgwACnI4mIVKuaHuaoB2Tvdy0bSNn/RmPMMCDGWjvrUG9qrX3WWtvXWts3KTkZgBgVOCL4/YGDZ/v168fdd9/NsmXLVNyISJ1Q0wVOHpC637VUILf8heBU1sPAjZX9gBJfYJLK5z+8gCLR4qOPPqJHjx5s3boVt9vNhAkTSE9PdzqWiEiNCLnAMcY0NcY8YIz5yhizyhhzdPD6WGNM3xDfZg0QY4w5qty1HsDK/e47CmgHfGmM2Q7MBJobY7YbY9od7APcwcXFxSW+ECOJRJeSkhLuuOMOTj/9dHw+H7m5uYf+IhGRKBPSGhxjTGfgCyAW+A44HkgIvtwJGEBgMfBBWWvzjTEzgYnGmD8BPYFzg19f3gqgfI/4AcAUoDe/nYd1QCU+P/FA+0bJh4ojEnV++eUXRo4cyYIFCxgzZgxPPPEESUlJTscSEalxoS4yfgTYAJxOYJrJU+61r6nELidgLPA8ga3fe4DrrLUrjTGDgA+ttfWstV5ge+kXGGMyAb+1dvsB37Gc0kZ/Hs1RSR00ceJEVq1axRtvvMFFF13kdBwREceEWuCcBFxmrc0yxuzfQW870DzUD7TWZgLnHeD6lwQWIR/oaz4DDtnkD8Ad7OmREKtGf1I3FBQUkJmZSatWrXjssceYMGECHTp0cDqWiIijKrNNvKJFLQ05xDbumlTaByc9KdbRHCI1YeXKlVx88cUkJCTw7bffkp6eroXEIiKEvsh4ERU04yPQuG9BeOJUXelhm+pkLNHMWsuzzz5L37592bVrFw888AAuNbcUESkT6gjO/cAcY8x7wH8IDJScaIy5BrgIOKWa8lWapfSoBn2zl+iUk5PDn/70J9566y2GDBnCyy+/TLNmzZyOJSISUUKqAqy1HxMoZHoArwIGeAw4C7jIWvt1tSWspLIRHDX6kygVGxvL2rVreeihh5gzZ46KGxGRAwh5DY61dqYxZhbQFWhCYAfUD9baiNquVFTiQ5tiJdr4/X7+9a9/kZGRQUpKCgsXLiQ2VuvMREQqUqmzqGzgqO4V1ZQlLOJiAoNSpSM5IrXd9u3bufzyy/noo4+w1nL99deruBEROYRQG/0dsqGGtfbNqseputLCRruoJBrMmzePjIwMcnJyeOaZZxgzZozTkUREaoVQR3Ber+B6+XGSiCpwtMhYartp06Zx9dVX07VrV/73v//RrVs3pyOJiNQaoRY4XQ5wrSFwNnABMCpsiaosuItKi4ylljv99NMZN24c999/v45bEBGppJAKHGvtTxW8NN8Y4wOuA74JW6oq8AdHcGJV4Egt9MYbb/DOO+8wffp02rRpw+OPP+50JBGRWikc8zifAueE4X3ConSKKs6toxqk9igoKGDMmDGMGDGCDRs2kJ2d7XQkEZFaLRwFTl+gIAzvExaljf4S47QGR2qHH374gX79+vHcc89x++2388UXX1C/fn2nY4mI1Gqh7qK67QCX44BuwDBgWjhDVUXpCI5bi4ylFvB6vQwbNoy8vDzmzp3LkCFDnI4kIhIVQl1k/OABrvmALcDjwL1hS1RFpdu6Sk8VF4lEWVlZJCcnExsby+uvv07r1q1p2rSp07FERKJGqMMciQf4FW+tbWetvd1aGzGniZfSAI5Eqm+++YaePXtyzz33ANC3b18VNyIiYXbIMsAYEwfcA3Sz1haX+xVRRzSUslaHbUpk8vv9TJ48mUGDBmGM4ZxzImZtvohI1DlkFWCt9QB/AZKrP074qL6RSLJ9+3ZOP/10xo8fz/Dhw1m2bBn9+/d3OpaISNQKtQxYDhxdnUHCRWtwJBJt27aNRYsWMW3aNF5//XXS0tKcjiQiEtVCXWR8G/CyMWattfbj6gxUVWVHNbg1hCPO8ng8vPfeewwfPpxevXqxceNGUlNTnY4lIlInhFoFPA+kA3ONMbnGmJ+NMWvK/aqo03GNK12DEx+jAkecs379ek444QQuuOAClixZAqDiRkSkBoU6grOYfQ/WjFilIWM1giMOef3117n66qtxu928/fbb9O7d2+lIIiJ1TqhnUY2o7iDhZAy4XVqDIzVv3LhxPPnkkwwYMIBXX32Vtm3bOh1JRKROqnCYwxiz3hjToybDhEustlCJQ/r168f48eP5/PPPVdyIiDjoYCM47YD4GsoRVjE6SVxqiLWWqVOnEh8fz+jRo7n00kudjiQiIoTnsM2IU+DxOR1B6oDMzEyGDx/O9ddfz+zZs8sWuIuIiPMOVeDUyu/YjerVyoEnqUW+/vprevbsyXvvvccjjzzCW2+9hVHvJRGRiHGoRcb3GmN2h/A+1lo7KhyBwiFGC4ylGm3YsIGTTz6ZNm3aMH/+fPr16+d0JBER2c+hCpyeQHEI7xNRIz3aQSXVoaioiISEBNq3b89LL73E2Wefrd42IiIR6lBTVOdZa9uH8KtDjaQNkRYZS7jNnj2b9u3bM3/+fAAuueQSFTciIhEsKhcZawRHwsXj8XDzzTdz1lln0aRJE+rXr+90JBERCUGonYxrFR20KeGwdu1aRo4cyaJFixg7diyPPPIIiYmJTscSEZEQRGWB8/POPKcjSBSYOXMma9euZebMmQwbNszpOCIiUgkVTlFZa13W2m9rMky4dGuptRFyePLy8soOx7z55ptZuXKlihsRkVooKtfguDRFJYdh2bJl9O3blzPPPJOCggLcbjctWrRwOpaIiByGqCxw1HBNKsNay5QpUzjuuOPIycnhtddeIykpyelYIiJSBVG5BkebqCRURUVFjBgxgnfeeYczzzyTF198kcaNGzsdS0REqig6R3CcDiC1Rnx8PMnJyTz22GO89957Km5ERKJElI7gqMSRivl8Ph566CEuvPBCjjrqKKZPn65pTRGRKBOVIzgqcKQiW7Zs4dRTT+XOO+/k1VdfBbRmS0QkGkXlCI5+XsmBfPDBB4waNYrCwkJeeOEFRo2KmPNhRUQkzDSCI3XCjBkzOPvss2nVqhWLFy/miiuu0MiNiEgUi8oCRz+3pJTf7wfgzDPPZNKkSSxYsIDOnTs7nEpERKpbVBY4Gvz1EAgAABr/SURBVMERgOnTp9OvXz9yc3NJTEzkzjvvJCEhwelYIiJSA6KywFF9U7fl5eUxatQoMjIySE5OJj8/3+lIIiJSw6KywNEITt21dOlS+vTpwyuvvMJdd93FJ598QrNmzZyOJSIiNSwqd1Gpk/H/t3fn4VVUaR7Hvy8hSFiVVWlFYABtbU2IiCIgpptWBn0EWlyQJf0gaONCI64MaLfKOIo22jAouCFIIy2jIqIgIiOytIz0sCij0AIqKDQGFAiGJeHMH1XRyyXbheTWvXV/n+epJ7lVp6ree0+SenPOqTqp68477yQ/P59FixZxySWXBB2OiIgEJJQJzp79hUGHIHG0c+dOnHM0atSIqVOnUrNmTRo1ahR0WCIiEqBQdlEdLDwcdAgSJx988AFZWVkMGTIEgFNPPVXJjYiIhDPBad5AM0GHXVFREQ888AA5OTnUrFmT0aNHBx2SiIgkkFB2UWmMcbht27aNvn37snjxYvr3789TTz1F3bp1gw5LREQSSCgTHN1FFW7Vq1dn27ZtTJ06lYEDBwYdjoiIJKBQdlHpLqrwOXDgAE888QSFhYU0btyYdevWKbkREZFShTTBUYYTJhs2bKBjx46MGDGCd955B/BacUREREoTygRHkyiGx7Rp08jOzubLL7/kjTfe4PLLLw86JBERSQKhTHDURRUOo0aNIjc3l/POO481a9Zw5ZVXBh2SiIgkiVC286uLKhyuuuoqatSowejRo0lLSws6HBERSSLhTHDUhJOUnHOMHz+ezZs38+STT5KdnU12dnbQYYmISBJSF5UkhLy8PK688kqGDx/O5s2bOXToUNAhiYhIEgtpgqMMJ5m8//77ZGZmsmDBAsaPH8/s2bNJT08POiwREUlioeyiKjysuaiSxe7du+nVqxdNmzZl7ty5tGvXLuiQREQkBEKZ4Hy3T90biS4vL4+GDRtSv3595s6dS1ZWFnXq1Ak6LBERCYlQdlG1aFQ76BCkDHPmzOGMM87g2WefBaBz585KbkREpFKFMsHREJzEtH//foYNG0bPnj1p0aIFOTk5QYckIiIhFfcEx8wamNnrZrbPzL40s+tLKXeXmX1iZnvNbLOZ3VXRc+guqsSzfv16OnbsyIQJExg+fDjLly+nTZs2QYclIiIhFcQYnInAQaApkAW8ZWZrnHProsoZMBBYC/wLsMDMtjjnZpZ3At1FlXg2btzI1q1befPNN7niiiuCDkdEREIuri04ZlYbuAq4zzmX75xbCswBBkSXdc6Ndc79r3Ou0Dm3HngD6FTB81Rm2HKM9u7dy9y5cwHo0aMHmzZtUnIjIiJxEe8uqrZAkXNuQ8S6NcDZZe1kXsbSBYhu5SmRuqiC9/e//53s7Gz69OnDtm3bAKhbt27AUYmISKqId4JTB9gdtW43UN6V7494sU4paaOZ3WhmK81sJYChDCcohw8fZty4cXTs2JH9+/fz7rvvcsoppwQdloiIpJh4j8HJB+pFrasH7C1tBzO7FW8sThfn3IGSyjjnngGeATjhlDZOLTjBcM7Ru3dv5syZQ8+ePXnhhRdo0KBB0GGJiEgKineCswGobmZtnHP/8NdlUkrXk5kNAu4FLnbOba3oSTTZZjDMjC5dunDppZdy8803ayyUiIgEJq4JjnNun5m9BjxoZoPx7qLqCVwUXdbM+gEPAznOuU2xnEfX1fgpLCzkwQcf5MILL6RHjx7ceeedQYckIiISyIP+bgYygB3Ay8BQ59w6M+tiZvkR5cYADYGPzCzfXyZV5AQagxMfW7ZsIScnh4ceeoj33nsv6HBERER+FPfn4DjndgG9Sli/BG8QcvHrlsd6DvVQVb3Zs2czaNAgDh06xPTp0+nXr1/QIYmIiPwolFM16EF/VWvp0qX07t2bVq1asWrVKiU3IiKScEKZ4Ci/qRoFBQUAdOrUiRdffJHly5fTunXrgKMSERE5WigTnLz8g0GHECrOOaZMmULLli35xz/+gZmRm5tLjRo1gg5NRESkRKFMcBrV0YW3suzZs4f+/fszaNAgzj77bGrXrh10SCIiIuUKZYKTnhbKtxV3K1euJDs7m7/+9a+MGTOGBQsW0KxZs6DDEhERKVcQs4lXOY3BqRxTpkzh4MGDLF68mE6dKjTPqYiISEJQU4cc4dtvv+XTTz8F4PHHH2f16tVKbkREJOmEMsFRA86xWbRoEZmZmVx77bUcPnyYjIwMzSUlIiJJKZQJjvqoYlNYWMh9991Ht27dqF+/PtOnT6datXD+aIiISGoI5xicoANIInl5efTq1Ytly5YxaNAgxo8frzulREQk6YUywZGKq1+/PnXr1mXGjBn07ds36HBEREQqRSj7IdRDVbaCggJGjhzJt99+S3p6Om+//baSGxERCZVwJjjqpCrVp59+ygUXXMAjjzzC22+/DYApIxQRkZAJZ4Kj6/VRnHM8//zztG/fnu3btzNv3jxyc3ODDktERKRKhDLBkaM98cQTDB48mI4dO7JmzRq6d+8edEgiIiJVJpSDjNWA85PDhw9TrVo1BgwYgJkxbNgw0tLSgg5LRESkSoWyBUddVF5i89hjj9GtWzcKCwtp3Lgxt99+u5IbERFJCeFMcFK8DWfHjh306NGDu+++mwYNGrB///6gQxIREYmrUCY4qZzfLFy4kMzMTBYvXsykSZOYNWsWderUCTosERGRuArlGJxUVVhYyC233MJJJ53EggULOOecc4IOSUREJBChTHBSrQHnq6++onHjxmRkZPDWW2/RrFkzatWqFXRYIiIigQllF1UqPbju1Vdf5dxzz2XkyJEAtG7dWsmNiIikvHAmOEEHEAcFBQUMHTqUPn36cMYZZzBs2LCgQxIREUkYoUxwwm79+vV06NCBSZMmcffdd7NkyRJatWoVdFgiIiIJI5xjcELehFOtWjUKCgqYP38+l112WdDhiIiIJJxQtuCEMcHZvXs348ePxzlHmzZt+Oyzz5TciIiIlCKcCU7IRuGsWLGCrKwsRowYwdq1awGoXj2UjW8iIiKVIpwJTkjym8OHDzN27Fg6d+6Mc44lS5aQmZkZdFgiIiIJT80ACWzAgAHMmDGDq6++mmeeeYYTTzwx6JBERESSghKcBNavXz+6du3KkCFDUurZPiIiIscrlAlOsiYDhw4d4v777+fEE0/knnvuoUePHkGHJCIikpTCOQYn6ACOwebNm7n44ot55JFH+Oqrr3DOBR2SiIhI0gplC06ymTVrFkOGDME5xyuvvMLVV18ddEgiIiJJLZwtOEnUhLNp0yb69u3LmWeeyerVq5XciIiIVIJQtuAkw3NwduzYQZMmTWjVqhXvvvsunTt3Jj09PeiwREREQkEtOHHmnGPy5Mm0bNmS+fPnA5CTk6PkRkREpBKFMsFJVN9//z3XXHMNv/vd7+jcuTPt2rULOiQREZFQCmWCk4gNOB9++CFZWVnMnj2bRx99lHnz5tG0adOgwxIREQmlUI7B2V9YFHQIR1mzZg1mxtKlS7nggguCDkdERCTUQtmCUyMtLegQANi+fTsLFy4E4MYbb+Tjjz9WciMiIhIHoWzBOaF68HnbO++8w8CBAwHvIX61atWiTp06AUclIiKSGoLPBKpAkHdRHTp0iHvuuYfu3bvTuHFjFi1aRK1atYILSEREJAWFsgUnqASnoKCAnJwcVqxYwU033cS4ceOU3IiIiAQgnC04Ad1HlZGRwcUXX8ysWbOYNGmSkhsREZGAhDLBiWd+88MPPzB06FBWrVoFwNixY+nTp0/8AhAREZGjhDLBiVd+8/HHH9O+fXsmT57MkiVL4nRWERERKU84E5wqHoTjnGPSpEl06NCBXbt2sWDBAoYNG1al5xQREZGKC2eCU8XHnzFjBkOHDqVr166sWbOGbt26VfEZRUREJBa6iyoGP/zwA7Vq1eKaa66hqKiI/v37U61aKHNEERGRpBbKq3Nl30VVVFTEww8/zM9//nPy8vJIT09n4MCBSm5EREQSlFpwyrFt2zb69+/PokWLuPbaa0lPT6+8g4uIiEiVCGeCU0nHmTdvHrm5ueTn5/Pcc88xaNCgKh/ALCIiIscvlAlOZWQ4zjkmTpzIySefzMyZMznrrLOO/6AiIiISF6FMcI5nDM7GjRupUaMGp512GtOmTSMjI4OMjIxKjE5ERESqWihHyR5rL9LLL79Mu3btGDp0KAANGjRQciMiIpKEwpngxFh+37593HDDDVx//fWcc845TJw4sUriEhERkfgIZ4ITQxPOxo0bad++PVOmTGHUqFEsXryY008/vQqjExERkaoWzjE4MTThNGnShCZNmjBx4kR++ctfVl1QIiIiEjfhbMEpZ/uuXbu44447KCgooG7durz//vtKbkREREIknAlOGRnO0qVLycrKYsKECSxbtswvr2fbiIiIhEkoE5yS2nCKiooYM2YMXbt2JT09nWXLlmmSTBERkZAKZYJTUoPM8OHDue+++7juuutYtWoV559/fvwDExERkbgI5yDjiO+LiopIS0vjtttu47zzziM3N1ddUiIiIiEXzgTHjIMHDzJy5Ei2bt3KzJkzadu2LW3btg06NBEREYmDuHdRmVkDM3vdzPaZ2Zdmdn0p5czMHjWznf4y1irY9LLli01cdNFFjBs3jsaNG1NUVFS5b0JEREQSWhAtOBOBg0BTIAt4y8zWOOfWRZW7EegFZAIOeBfYBEwq6+BFBXu57l+7ckKNdF577TV69+5d6W9AREREEps55+J3MrPawHfAL5xzG/x1LwFfO+fujSq7HHjROfeM//oGYIhz7sIyz1EtzbVrfwGz/2smzZs3r5L3ISIiIlXLzP7unGt/rPvHuwWnLVBUnNz41gBdSyh7tr8tstzZJR3UzG7Ea/EBOLDqo799oukWEkYjIC/oIORHqo/EovpILKqPxHLG8ewc7wSnDrA7at1uoG4Fyu4G6piZuahmJ7+Vp7ilZ+XxZHxSuVQfiUX1kVhUH4lF9ZFYzGzl8ewf70HG+UC9qHX1gL0VKFsPyI9ObkRERESixTvB2QBUN7M2EesygegBxvjrMitQTkREROQIcU1wnHP7gNeAB82stpl1AnoCL5VQfBowwsx+ZmbNgDuAFytwmmcqK16pFKqPxKL6SCyqj8Si+kgsx1Ufcb2LCrzn4AAvAL8GdgL3OudmmFkXYJ5zro5fzoBHgcH+rs8B96iLSkRERMoT9wRHREREpKqFcrJNERERSW1KcERERCR0kjLBicd8VlJxMdTHXWb2iZntNbPNZnZXvGNNBRWtj4jyNczsMzPbGq8YU0UsdWFm2Wb2gZnlm9k/zez38Yw1FcTwt+oEM5vk18MuM3vTzH4W73jDzsxuNbOVZnbAzF4sp+ztZrbdzHab2QtmdkJ5x0/KBIcj57PqBzxtZiU95ThyPqtzgSuAm+IVZAqpaH0YMBA4CegO3Gpm18UtytRR0foodhewIx6BpaAK1YWZNQLmA5OBhkBrYEEc40wVFf3d+D3QEe+60Qz4HpgQryBTyDfAGLwbj0plZpcB9wK/AloArYAHyjt40g0yjsd8VlJxsdRHCfuOx/sZvK3qI00NsdaHmbUE3gZGAM86506NZ7xhFuPfqoeB05xzA+IfaWqIsT6eBvY65+72X18OjHPOHdfUAVIyMxsDnOqc+20p22cAXzjn/s1//SvgL865k8s6bjK24JQ2n1VJWXiF57OSYxZLffzI7yrsgh7eWNlirY8JwL8BBVUdWAqKpS4uBHaZ2XIz2+F3iWi24MoVS308D3Qys2ZmVguvtWdeHGKUkpV0LW9qZg3L2ikZE5xKmc+qimJLRbHUR6Q/4v38TamCmFJZhevDzHoD1Z1zr8cjsBQUy+/GqUAuXtdIc2Az8HKVRpd6YqmPDcBXwNfAHuDnwINVGp2UpaRrOZRznUnGBEfzWSWWWOoD8AaW4Y3Fudw5d6AKY0tFFaoPv7l+LKDuwaoTy+9GAfC6c+4j59x+vPEFF5lZ/SqOMZXEUh9PAzXxxkPVxnsCv1pwglPStRzKuM5AciY4ms8qscRSH5jZIPzBYs453bVT+SpaH23wBustMbPteH/AT/HvUmgRhzhTQSy/G2uByH+8ir9Xa3PliaU+MvHGb+7y/wmbAHTwB4NL/JV0Lf+nc25nWTslXYITp/mspIJiqQ8z6wc8DPzaObcpvpGmhhjq4xPgNCDLXwYD//S/3xK/iMMrxr9VU4DeZpZlZunAfcBS59z38Ys43GKsj4+AgWZW36+Pm4FvnHN58Ys4/MysupnVBNKANDOraWbVSyg6DbjBzM4ys5OA0VTkWu6cS7oFaADMBvbh9ZNe76/vgtcFVVzO8Jrhd/nLWPw7x7QEUh+bgUN4zY3Fy6Sg4w/bUtH6iNrnEmBr0LGHbYmlLoCheGM+vgPexLurKvD3EKYlhr9VDYG/4D0+4XtgKdAh6PjDtuCNxXRRyx/xxqHlA80jyo7A+ydsD94/BCeUd/yku01cREREpDxJ10UlIiIiUh4lOCIiIhI6SnBEREQkdJTgiIiISOgowREREZHQUYIjIiIioaMERySBmNlvzcyVsnSL8ViD/f3iMkO4mY2Jivc7M1thZtdVwbmq++cYHbHuN2Y2vISy3fyynSs7jjLiax31WRSZ2TYze8nMfnaMx8w2sz+a2YmVHa9IGJX0xEARCd7VQPRUFv8XRCDHoKP/tSFwE/CymdVwzk2rrBM45wrNrCNHPnX5N0Bn4Mmo4v/jxxTENC1jgLeAE/wY7gfONLOOzrnCGI+VDfwB7wmuesKxSDmU4IgkptXOuc+DDuJYOOc+LP7ezBYA64HheI9br5LzlFNuD1ChslVgY0Sci83sBLwntWYBKwOKSSQlqItKJMmYWYaZ/dnM1pnZPr/rY46ZnVGBfQeY2Wp/v91mttbMBkeVyTGzRWaW7y/zzOysY4nVOXcIWA20jjh+fTN7yo/7oJmtN7PfR8VQz8z+08y2mNkBM/unmb1rZm397Ud0UZnZdKAfcHpEt9Dn/rYjuqjM7Bkz+8bM0qLOWdP/TB6PWNfEzCb75Q+a2admdsOxfBa+//W/No869xgzW2Vme8wsz8zeM7MOEdsHA8/6LzdHvMdTIz6PUf5necDMvjazx/yESiQlqQVHJDGlRU0655xzRf73Gf7yILAdryvoFuBvZnamc25HSQc0s67AVLwunDvwJrg7CzgpokxP4FVgDnA93j9B9+LNOn6uc+7rY3gvLfG7VPykYh5wLt6EkuuAK4Enzayhc+5+f58/A92BUcDnQCO87qf6pZzjD36ZTKC3v25/KWWnAUOAXwELItb3BOrhT77oj3VZBqTjdS19AfQAnvW73J6u0Ls/Ugv/68ao9c2AP+F1S9YBcvE+82zn3DrgDaAVMBKvK26bv19xXb8M/CvwCF5r1dl4Px/NgWuPIU6R5Bf0ZFtatGj5aQF+y9GTzzm8maVL2ycNqA38ANwWsX6wv++p/ut7gR1lHMfwLuLvRK0/EW+y2sfLiX2Mf77q/tIUeMhf97hfppf/un/Uvi/iJSQN/NefAWPLOFd1/zijI9ZNB74ooWw3v2zniPe5CXgpqtxcYG3E6weAAuBfospNwZv0L62M+Fr75xzkx1obL6H6BphZzueYhpdUbQT+VEJ9togqn+Ovvz5qfa6//pygf661aAliUReVSGLqDZwfsRzRLWJm15nZ/5jZbqAQb+bdDKCsbqqPgMZmNs3MLjez6NaQM4HTgb/4XR7V/VakfGAFcHEFYz/kL9uBu4BxeC0x+McoBGZG7TMdbyDuBRGx3mBm95rZeWZWaX+rnHPOP19vM6sNYGaNgcs4cpxQd2A58GXU5/EO0ISyP+tiz+N9FvnAQrwWmtzoQmZ2qZm9b2Y78T6fg3gtNhU5R3e85PD1qDiLW6e6VOAYIqGjBEckMX3inFsZsawv3mBmvfG6JD4B+uIlBefjtbLULO2Azrn38LorWgCzgTwzW2Bmv/CLNPG/TuWnJKV46Y7XFVYRxUlZa6Cuc+4O59wBf1sDIM8dfQfR9ojtADfjjTkZgjcYd4eZ/cnMMioYQ3mm4bWq/MZ/3Rfv7+GMiDJNgF9y9Gfxsr+9Ip/HA3ifxSXA0/73EyILmNn5eHda7cZr8bnQL/cJZdRnVJw18VrwIuP8JoY4RUJHY3BEks91wGfOuUHFK8ysJl5XUpmcc68Ar5hZHbyL96PAPDNrDuz0i90N/HcJux8oYV1J5yjr7qBdQCMzqx6V5Jzsf93pH2MvXpfavWbWAu+2+f/Aa6kYxXFyzn1uZh8C/fHG3PQH3nPOfRNRbCfebegjSjnM+lLWR/oi4vNYbGb1gMFmNsk5VzzguA/e+7oq8jMxswZ4XWHl2YmX3HQtZfs3pawXCTUlOCLJpxZeN0akgcTQIuucywfmmFlrvMGtJ+E9Z2cLcJZz7rFKijXaYuB24CrgrxHr++Fd5FeUEOsXwGNmNgD4RfT2CAfwuukq6iVgvJnl4LWYDIjaPh/vOT5fOOfyYjhuWe7Be+9/wBvUDD/VpysuZGaX4g08/jRi3+IEM/o9zscbNF7bObe4kuIUSXpKcESSz3zgP/3bmefhXZxvAfaUtZOZ/Tted8V/492F0xy4FVjpnNvll7kVeM1vEZqF1zpwMnARsMk59+fjjH0u8De8O5FOxruAX4E3uPoh59x3fhwrgNfwumn24Q2kPRuYXMax/w8YZGY3AquAAufcJ2WUnwk8gZfo7ANej9r+OF7L0RIzewLYANTFG6t0kXOuNzFyzn1tZpOA4WaW5ZxbjVeftwJTzGyqf/zRHN3yUvygx1v92+IPAWuccwvNbBbeGJxxeA82BK8rsgdwh3Mu+q4tkfALepSzFi1aflr46S6q1mWUSQMexrsA/oCXsGTiDWB9LqJc9F1UV+INPN2G1xqwBW+cy8lRx++ENybkO7xWlc14404uLCf2MfhjeMspVx94yo/jIF5Xz++jyjyOl6Tsxhuguxa4NWJ7SXdR1cVrFfrO3/a5v/6Iu6iizvO6v21aKbE2wLtl/Qs/1h3AB0TcrVbKfsV3Uf22hG1N/Pf0asS64f45CvASlBxgKbAwat8H/XoviqrbNLyWsbV+nX2P9/yhR4F6Qf9ca9ESxGLO/dgqKiIiIhIKuotKREREQkcJjoiIiISOEhwREREJHSU4IiIiEjpKcERERCR0lOCIiIhI6CjBERERkdBRgiMiIiKh8//cb6kOYfXB+AAAAABJRU5ErkJggg==\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--') #clasificador aleatorio\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.xlabel('False Positive Rate', fontsize=16)\n",
    "    plt.ylabel('True Positive Rate', fontsize=16)\n",
    "\n",
    "plt.figure(figsize=(8, 6))\n",
    "plot_roc_curve(fpr, tpr)\n",
    "save_fig(\"roc_curve_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9563121591738021"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score\n",
    "\n",
    "roc_auc_score(y_train_5, y_scores)#clasificador perfecto AUC = 1; clasificador aleatorio AUC = 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "forest_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
    "y_probas_forest = cross_val_predict(forest_clf, X_train, y_train_5, cv=3,\n",
    "                                    method=\"predict_proba\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_scores_forest = y_probas_forest[:, 1] # score = proba of positive class\n",
    "fpr_forest, tpr_forest, thresholds_forest = roc_curve(y_train_5,y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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": [
    "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.legend(loc=\"lower right\", fontsize=16)\n",
    "save_fig(\"roc_curve_comparison_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Se recomienda usar la curva precisión vs exahustividad cuando la clase positiva es escasa o cuando es más importante en la aplicación evitar los falsos positivos que los falsos negativos, y la curva ROC en caso contrario (es más importante evitar los falsos negativos que los falsos positivos, ej. problemas de salud)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.992830497450426"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roc_auc_score(y_train_5, y_scores_forest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9806437581557199"
      ]
     },
     "execution_count": 92,
     "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": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.831765356945213"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "recall_score(y_train_5, y_train_pred_forest) #pocos 5 recall = TP/(TP+FN), de los pocos 5s se equivoco, se sesga a clasificar no 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clasificación multiclase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=int8)"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.fit(X_train, y_train)\n",
    "sgd_clf.predict([some_digit]) # si el clasificador es netamente binario aplica estrategia one vs all, \n",
    "#a excepción de las svm que aplica one vs one para evitar costo computacional"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-182692.06194962, -401178.58857685, -342030.93878097,\n",
       "         -89340.70455522, -512749.92629109,   75090.39935667,\n",
       "        -674198.61305684, -310540.47612933, -572399.68212808,\n",
       "        -387233.04985949]])"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "some_digit_scores = sgd_clf.decision_function([some_digit])\n",
    "some_digit_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(some_digit_scores) # máximo score para 5 vs all"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], dtype=int8)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.classes_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sgd_clf.classes_[5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=int8)"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.multiclass import OneVsOneClassifier #extensión a one vs one\n",
    "ovo_clf = OneVsOneClassifier(SGDClassifier(max_iter=5, tol=-np.infty, random_state=42))\n",
    "ovo_clf.fit(X_train, y_train)\n",
    "ovo_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "45"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(ovo_clf.estimators_) #se generan 45 combinaciones C(C-1)/2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([5], dtype=int8)"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf.fit(X_train, y_train)\n",
    "forest_clf.predict([some_digit]) #algunos clasificadores ya incluyen multiclase desde el score o probabilidad de membresia"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0. , 0. , 0. , 0. , 0. , 0.8, 0. , 0.2, 0. , 0. ]])"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "forest_clf.predict_proba([some_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.87482503, 0.88729436, 0.86032905])"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(sgd_clf, X_train, y_train, cv=3, scoring=\"accuracy\") # un clasificador aleatorio obtendria un 0.1 dado que C = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.90496901, 0.91339567, 0.9099865 ])"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import StandardScaler #el preproceso puede ayudar a mejorar la clasificación\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": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[5727,    3,   28,    9,   10,   49,   47,    9,   37,    4],\n",
       "       [   1, 6450,   54,   28,    6,   48,    5,    8,  130,   12],\n",
       "       [  57,   35, 5342,  102,   77,   25,   80,   52,  171,   17],\n",
       "       [  57,   38,  127, 5336,    2,  244,   38,   58,  131,  100],\n",
       "       [  20,   29,   41,    8, 5354,    9,   52,   33,   77,  219],\n",
       "       [  68,   44,   33,  193,   72, 4610,  120,   26,  161,   94],\n",
       "       [  34,   22,   50,    2,   37,   85, 5645,    5,   38,    0],\n",
       "       [  23,   23,   74,   28,   50,   13,    6, 5806,   16,  226],\n",
       "       [  52,  160,   78,  156,   15,  164,   61,   24, 4989,  152],\n",
       "       [  45,   34,   28,   93,  138,   37,    2,  198,   66, 5308]],\n",
       "      dtype=int64)"
      ]
     },
     "execution_count": 110,
     "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": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(matrix):\n",
    "    \"\"\"If you prefer color and a colorbar\"\"\"\n",
    "    fig = plt.figure(figsize=(8,8))\n",
    "    ax = fig.add_subplot(111)\n",
    "    cax = ax.matshow(matrix)\n",
    "    fig.colorbar(cax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure confusion_matrix_plot_barra\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(conf_mx) # matriz de confusión con barra de colores\n",
    "save_fig(\"confusion_matrix_plot_barra\", tight_layout=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "row_sums = conf_mx.sum(axis=1, keepdims=True)\n",
    "norm_conf_mx = conf_mx / row_sums #normalizar por muestras por clase para calcular como porcentaje"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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) # resaltar errores, diagonal se omite\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": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure error_analysis_digits_plot\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cl_a, cl_b = 3, 5 #analizar errores puntuales\n",
    "X_aa = X_train[(y_train == cl_a) & (y_train_pred == cl_a)] # 3 como 3\n",
    "X_ab = X_train[(y_train == cl_a) & (y_train_pred == cl_b)] # 3 como 5\n",
    "X_ba = X_train[(y_train == cl_b) & (y_train_pred == cl_a)] # 5 como 3\n",
    "X_bb = X_train[(y_train == cl_b) & (y_train_pred == cl_b)] # 5 como 5\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",
    "save_fig(\"error_analysis_digits_plot\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clasificación multi-etiqueta (cada etiqueta es binaria)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=None, n_neighbors=5, p=2,\n",
       "           weights='uniform')"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "y_train_large = (y_train >= 7) # digito mayor o igual a 7\n",
    "y_train_odd = (y_train % 2 == 1) # digito impar\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,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[False,  True]])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_clf.predict([some_digit])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Warning**:La siguiente celda puede ser costosa de calcular (puede tardar horas depiendo del hardware)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.97709078477525"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_knn_pred = cross_val_predict(knn_clf, X_train, y_multilabel, cv=3, n_jobs=-1)\n",
    "f1_score(y_multilabel, y_train_knn_pred, average=\"macro\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Clasificación multi-salida -> extensión de multietiqueta a multicalse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "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": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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 = 5500\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()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving figure 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\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Material extra"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dummy clasificador aleatorio"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.dummy import DummyClassifier\n",
    "dmy_clf = DummyClassifier()\n",
    "y_probas_dmy = cross_val_predict(dmy_clf, X_train, y_train_5, cv=3, method=\"predict_proba\")\n",
    "y_scores_dmy = y_probas_dmy[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "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": [
    "fprr, tprr, thresholdsr = roc_curve(y_train_5, y_scores_dmy)\n",
    "plot_roc_curve(fprr, tprr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clasificador KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=-1, n_neighbors=4, p=2,\n",
       "           weights='distance')"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "knn_clf = KNeighborsClassifier(n_jobs=-1, weights='distance', n_neighbors=4)\n",
    "knn_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_knn_pred = knn_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9714"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "accuracy_score(y_test, y_knn_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "from scipy.ndimage.interpolation import shift\n",
    "def shift_digit(digit_array, dx, dy, new=0):\n",
    "    return shift(digit_array.reshape(28, 28), [dy, dx], cval=new).reshape(784)\n",
    "\n",
    "plot_digit(shift_digit(some_digit, 5, 1, new=100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((300000, 784), (300000,))"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_expanded = [X_train]\n",
    "y_train_expanded = [y_train]\n",
    "for dx, dy in ((1, 0), (-1, 0), (0, 1), (0, -1)):\n",
    "    shifted_images = np.apply_along_axis(shift_digit, axis=1, arr=X_train, dx=dx, dy=dy)\n",
    "    X_train_expanded.append(shifted_images)\n",
    "    y_train_expanded.append(y_train)\n",
    "\n",
    "X_train_expanded = np.concatenate(X_train_expanded)\n",
    "y_train_expanded = np.concatenate(y_train_expanded)\n",
    "X_train_expanded.shape, y_train_expanded.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
       "           metric_params=None, n_jobs=-1, n_neighbors=4, p=2,\n",
       "           weights='distance')"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn_clf.fit(X_train_expanded, y_train_expanded)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_knn_expanded_pred = knn_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9763"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_test, y_knn_expanded_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.       , 0.       , 0.5053645, 0.       , 0.       , 0.       ,\n",
       "        0.       , 0.4946355, 0.       , 0.       ]])"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ambiguous_digit = X_test[2589]\n",
    "knn_clf.predict_proba([ambiguous_digit])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "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": [
    "plot_digit(ambiguous_digit)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 94.90%\n",
      "Recall: 97.89%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "\n",
    "X_test_transformed = preprocess_pipeline.transform(X_test)\n",
    "\n",
    "log_clf = LogisticRegression(solver=\"liblinear\", random_state=42)\n",
    "log_clf.fit(X_train_transformed, y_train)\n",
    "\n",
    "y_pred = log_clf.predict(X_test_transformed)\n",
    "\n",
    "print(\"Precision: {:.2f}%\".format(100 * precision_score(y_test, y_pred)))\n",
    "print(\"Recall: {:.2f}%\".format(100 * recall_score(y_test, y_pred)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.1"
  },
  "nav_menu": {},
  "toc": {
   "navigate_menu": true,
   "number_sections": true,
   "sideBar": true,
   "threshold": 6,
   "toc_cell": false,
   "toc_section_display": "block",
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}