tp_ml.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_svmlight_file\n",
    "import numpy as np\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "data = load_svmlight_file(\"batch.dat\") #charger la base de données dans le format svmlight / libsvm en une matrice creuse\n",
    "#Le premier élément de chaque ligne va étre utlisé comme cible pour y faire de la classification .\n",
    "x=data[0].todense() #retourne une représention dense de la matrice\n",
    "y=data[1] #la variable cible à prédire est la classe à laquelle un élement appartient\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tuple"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.matrixlib.defmatrix.matrix"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(x) #x est un tableau numpy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1244, 129)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.shape(x) #1244 observations et 129 attributs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1244,)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.shape(y) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(<1244x129 sparse matrix of type '<class 'numpy.float64'>'\n",
       " \twith 160476 stored elements in Compressed Sparse Row format>,\n",
       " array([1., 1., 1., ..., 5., 5., 5.]))"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "numpy.ndarray"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "matrix([[ 1.00000000e+01,  2.17495635e+04,  2.62748800e+00,\n",
       "          3.50752400e+00,  4.11089900e+00,  5.01569700e+00,\n",
       "         -3.47473800e+00, -4.56969300e+00, -1.34484910e+01,\n",
       "          2.11380440e+04,  2.71328100e+00,  3.61221300e+00,\n",
       "          4.28157000e+00,  5.02698600e+00, -3.25212700e+00,\n",
       "         -3.99553000e+00, -1.15995780e+01,  2.60182490e+03,\n",
       "          2.08574800e+00,  7.23314000e-01,  1.16123700e+00,\n",
       "          1.58737600e+00, -5.16412000e-01, -8.24780000e-01,\n",
       "         -1.11288200e+00,  2.60268650e+03,  2.07497700e+00,\n",
       "          7.23795000e-01,  1.21627300e+00,  2.59486600e+00,\n",
       "         -5.20201000e-01, -8.55015000e-01, -2.79124300e+00,\n",
       "          1.21854970e+03,  2.09254400e+00,  3.95898000e-01,\n",
       "          8.63491000e-01,  1.14535800e+00, -2.52613000e-01,\n",
       "         -4.08573000e-01, -1.15076800e+00,  1.40778760e+03,\n",
       "          2.10446300e+00,  4.46390000e-01,  9.12604000e-01,\n",
       "          1.21873600e+00, -2.92121000e-01, -4.84320000e-01,\n",
       "         -1.30983900e+00,  3.32170900e+03,  2.41484000e+00,\n",
       "          1.03039000e+00,  1.95922800e+00,  2.46039500e+00,\n",
       "         -7.09820000e-01, -1.18497500e+00, -1.38716300e+00,\n",
       "          3.50109450e+03,  2.40644200e+00,  1.06871200e+00,\n",
       "          1.98429500e+00,  2.35454200e+00, -7.44714000e-01,\n",
       "         -1.22730900e+00, -3.47117200e+00,  2.04074551e+04,\n",
       "          2.84916300e+00,  3.24659700e+00,  3.79411500e+00,\n",
       "          4.48536200e+00, -2.90292300e+00, -3.66601700e+00,\n",
       "         -9.83938800e+00,  2.47739990e+04,  2.94799500e+00,\n",
       "          3.89824300e+00,  4.52610700e+00,  5.47442000e+00,\n",
       "         -3.69141900e+00, -4.55346700e+00, -1.20509950e+01,\n",
       "          3.17388400e+03,  2.20168500e+00,  8.71304000e-01,\n",
       "          1.31597100e+00,  1.62918000e+00, -6.04284000e-01,\n",
       "         -9.11772000e-01, -2.62225500e+00,  2.36901440e+03,\n",
       "          2.20292700e+00,  6.75708000e-01,  1.11045200e+00,\n",
       "          1.40552600e+00, -4.77064000e-01, -7.60527000e-01,\n",
       "         -9.86015000e-01,  1.63615880e+03,  2.16398000e+00,\n",
       "          5.35716000e-01,  1.05298600e+00,  1.82069600e+00,\n",
       "         -3.32063000e-01, -5.30287000e-01, -1.50084100e+00,\n",
       "          1.65730800e+03,  2.16184400e+00,  5.53001000e-01,\n",
       "          1.15784600e+00,  1.50813300e+00, -3.34765000e-01,\n",
       "         -5.35050000e-01, -1.52105100e+00,  3.95540940e+03,\n",
       "          2.51981500e+00,  1.20570500e+00,  2.14597700e+00,\n",
       "          2.59708200e+00, -8.38715000e-01, -1.35217900e+00,\n",
       "         -3.79871500e+00,  3.57304470e+03,  2.58215900e+00,\n",
       "          1.12036400e+00,  2.11335800e+00,  2.38800700e+00,\n",
       "         -7.60981000e-01, -1.21810400e+00, -1.60570400e+00]])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x[0] #affiche la première classe"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Considèrons maintenant qu'on ne connait pas la classe pour chaque observation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DescribeResult(nobs=160476, minmax=(-12826.543, 316646.0625), mean=4676.709337064957, variance=552122996.7785624, skewness=7.904085291870655, kurtosis=71.4678053910924)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#mesures statistiques\n",
    "stats.describe(x, axis=None) #on met axis à None pour calculer les statistiques sur la totalité de la matrice sans prendre en compte les classes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Les valeurs sont distibuées loins de leur moyenne (variance très grande)\n",
    "En total, et sans prendre en compte les différentes classes, il y a 160476 individus (valeurs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "scipy.stats.stats.DescribeResult"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(stats.describe(x, axis=None))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DescribeResult(nobs=1244, minmax=(1.0, 6.0), mean=3.422829581993569, variance=2.4534124214572666, skewness=-0.23213972614475584, kurtosis=-1.5891191879204132)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stats.describe(y, axis=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "En totale il y a 6 classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#visualisation des mesures statistiques"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'whiskers': [<matplotlib.lines.Line2D at 0x1250f66a0>,\n",
       "  <matplotlib.lines.Line2D at 0x1250f6c88>],\n",
       " 'caps': [<matplotlib.lines.Line2D at 0x1250d7b38>,\n",
       "  <matplotlib.lines.Line2D at 0x126731dd8>],\n",
       " 'boxes': [<matplotlib.lines.Line2D at 0x1250f6978>],\n",
       " 'medians': [<matplotlib.lines.Line2D at 0x126731a90>],\n",
       " 'fliers': [],\n",
       " 'means': []}"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "fig4, ax4 = plt.subplots()\n",
    "ax4.set_title('Boite à moustache des données')\n",
    "ax4.boxplot(A, showfliers=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [],
   "source": [
    "#normalisation des données\n",
    "from sklearn.preprocessing import StandardScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "StandardScaler(copy=True, with_mean=True, with_std=True)"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scaler=StandardScaler()\n",
    "scaler.fit(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaled_data=scaler.transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Réduction de dimension"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.decomposition import PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "PCA(copy=True, iterated_power='auto', n_components=4, random_state=None,\n",
       "  svd_solver='auto', tol=0.0, whiten=False)"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acp=PCA(n_components=4)\n",
    "acp.fit(scaled_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.42199588 0.23057832 0.08285719 0.06440313]\n"
     ]
    }
   ],
   "source": [
    "print(acp.explained_variance_ratio_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[260.23107078 192.3597851  115.31083929 101.66197441]\n"
     ]
    }
   ],
   "source": [
    "print(acp.singular_values_) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_acp=acp.fit_transform(scaled_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "blues = y == 1\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '2em coposante (23.05%)')"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAClCAYAAABfoUaZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJztnXl8XGX1/9+fSdo0Gbrve8pO2UoTQFSgUBAVWURBUhBQ1kYsFv3+wK8KWPgqLoCCVZYibdNCBYqIgChlEb/K1kIpawFpbcvW8oUWSEubZM7vj+eZZjKdmdxJJkuT5/163dfcuctzz71z59xzz3Oec2RmBAKBQKDrE+toAQKBQCDQPgSFHwgEAt2EoPADgUCgmxAUfiAQCHQTgsIPBAKBbkJQ+IFAINBNCAq/iyPpvyXN2l7azXCcckkmqbgF+14maV5byBUIbI8Ehd8JkbRS0iZJH0t6V9JsSTu0pC0z+4mZndVKeSZJWlPodtsT/+BYmWH5o5I+kFSSYd0Bku6XtF7S+5KekvQNv26SpIT/jT6StDy5LhDorASF33k5xsx2ACYClcAP0zeQI/yGLURSOXAwYMCxaesOAh4G/g7sDAwEpgJfSNnsLf8b9QEuAm6SNL7NBQ8EWkhQFp0cM3sT+AuwF2y1SP9H0j+BjcCOkkZIusdboa9LOju5f7pbQ9KnJP3LW63PSZqUsm6ApFskveWt3rslxf3xR3hr9mN/vPR2j5X0om/3UUl7pKxbKel7kpZJ2iDpD5J6ZTpfSUWSfinpPUlvAEenre8r6WZJb0t6U9IVkopaeHlPA54AZgOnp637BTDHzH5mZu+ZY4mZnZTeiF93N/ABsI3Cl9Rf0r2S1vnreq+kUX7d1yQtTtt+uqR7/HyJvx6r/Nve9ZJKU7Y9TtJSSR9K+rekz2c60eZ+A0ln+3vnfX8vjfDLZ0q6Kq2teyRN9/MjJC3057ZC0rSU7Q6Q9Li/J96W9BtJPf26bVx1/r45y8//TtLClHU/k/SQJGU6v0BEzCxMnWwCVgJH+PnRwIvA5f77o8AqYE+gGOgBPAb8FugFTADWAYf77S8D5vn5kcD/AV/EPeyP9N8H+/X3AX8A+vt2D/XLJwFr0mRMbXdXoNa31wP4f8DrQM+U83kKGAEMAF4Gzsty7ucBr/jzHgA8grPAi/36PwI3AHFgiG/33CxtbZUxy/rXgWqgAqgDhvrlZUADcFiOfbdeE38tv+zb2C3DtgOBr/h2ewN3AHenHOsjYJeU7Z8GTvbz1wD3+GvRG/gz8FO/7gBgg7/uMf/77p7jnsr4GwCHA+/h3iZLgOuAx1KO8RYQ898H4QyNof6YS4BLgJ7AjsAbwFF+2wrgU7j7tNwf8zt+XXnq75pyb5+Vcl1eBc7AvYW9B4zq6P/m9j51uAAZbszfA2uBFyJsewjwDFAPfDVt3QPAeuDejj6nFlyDlcDHXv7/4JR5qV/3KDAjZdvRXjn1Tln2U2C2n7+MRsV8EVCTdqy/4qzb4UAC6J9BnknkVvg/Am5PWRcD3gQmpZzPqSnrfw5cn+XcHyblYQB8LqkYvJLZnLwWfn0V8EiWtrbKmGHdZ3EKepD//gow3c+P9MfMqDxTrknC/0bvA0vxSjrC7zsB+CDl+zzgEj+/C+4BUAYI9yDdKWXbg4AVfv4G4Jo87qmMvwFwM/DzlHU7+GtT7r+/DBzp588H7vfzBwKr0o7zfeCWLDJ8B/ijny8nh8JPaf993H+gqqP+j11p6owundlAxtfSDKzCWQC3Zlj3C+DrhRGpQzjezPqZ2VgzqzazTSnrVqfMjwDeN7OPUpb9B6e00hkLnOhfsddLWo9TfMNxD473zeyDFsg6wh8TADNLeBlTZXgnZX4jTqlkayv1/P6TMj8W9wbxdor8N+As/Xw5Hfibmb3nv99Ko1vnA5wyH95MG2/532iAmU0wswWZNpJUJukGSf+R9CHujaxfiivqVtyDC2AKzvrfCAzGKf4lKef7gF8O7jf7dx7nnO03SP/9Psa9+SV/vznAqX7+VKDGz4/FufpS76f/xj2YkbSrd1+948/7J7g3hEiY2ZO4NwYBt0c+y0BWcoa6eR/fl3CvVCOATcALwH1m9mJbCGRmj8l1pqXKsRMwE3ejbwTONrNXzGylX5/I0M5Dqf7pLkZqitO3gAGSeqco/TE4Czud1TgL/+z0FZKG+3b6mdn6HMfLxFvA3iltCaeMMsnQHG/7fZOMSZlfjbPwB5lZfQvaTspXCpwEFElKKsESnBLe18yek/Q4zg3zSEuPk8J3gd2AA83sHUkTgGdxigzgQWCwX14FTPfL38P95/Y015eTzmpgpwLI9xZOeQPg+20G0vj7zQNekLQvsAdwd8rxV5jZLlna/R3uPKvM7CNJ3wG+6tfV+s8y4EM/Pyx1Z0nfwv0ub+HchD9t0dkFtpLVwpf0Y+CfuFfIJ3GW1O0498mVkh6UtE+7SAk3At82swrgezgXRwAws9XAv4CfSurlf5MzcX/SdOYBx0g6yneO9pILLxxlZm/jOmd/6zsZe0g6xO/3LjBQUt8sYtwOHC1psqQeOAW32cuVL7cD0ySNktQfuDjlXN8G/gZcJamPpJiknSQdmucxjse5wcbj3CsTcIrsH7iOXHAK5gxJ/yVpIICkfSVltOKboTdOca+XNAC4NHWlmdXh/Pq/wPnXH/TLE8BNwDWShngZRko6yu96M/ANf91jft3uLZDvNt/OBLnw1J8ATyYNKjNbg+tXqAEWprxtPgV8JOkiSaX+ntpL0v4p5/0h8LGXa2rKOa/DPVBO9ft9k5SHl6RdgStwbxRfB/6ffyAGWkEul85TZlZhZt81s1vNbJGZ3WtmV5vZMcApuI6aNkUu/vzTwB2SluIePM29anc3qnA+0bdwnZqXmtmi9I38w+E43Gv3OpyF9l803gdfx/luX8H1o3zH7/cKTim84V/dR6S1uxz3x7wOZ5Uegwsr3dKCc7kJ16/wHK5/5q609afh7ruXcK6XO8n/fjgd52deZWbvJCfgN8ApkorN7F+4zszDcef9Ps7wuL8F5/QroBR3bZ7AuWXSuRU4Argj7e3lIlzn8hPeLbII97aAmT0FfAPXsbsBF0I6ljzx98qPgIW4N6ydgJPTNpuDe4urSdmvAecBmACs8Oc3C0gaBt/Duag+wv2uf0hr82zc/fd/uCCEfwH4yJ15wM/M7Dkzew13z9Yow3iJQHRkFr0Ainfx9DSzD5vduBV4l869ZraXpD7AcjPL+qeWNNtvf2fa8knA98zsS20nbedG0gxcdMM3O1qWwPaLf9ubB4y1fJRGoFMRudPWx8feDSyU1G6+NP9wWSHpRC+HvC8x0Azelz4eZ30FAi3Cu+kuAGYFZb99k8uHf2zaoiPM7PNmdiQujrtNkHQb8Diwm6Q1ks7EuY/OlPQcLib9OL/t/nJD/k8EbpD0Yko7/8D5RSf7do5KP1Y34BlgFO51OhDIG7kBdOtxbrNfdbA4gVaS1aUj6Qe4QReXmtlSSf+NG1hhQNzMprSfmIFAIBBoLTl9+JKGATNw4WM/wvW6l5rZsvYRLxAIBAKFojmF3xsXvrYLcDmwGDci75P2ES8QCAQChSLrwCtJV+BcOsXAPWZ2rPfr3y9ptpnNLbQwgwYNsvLy8kI3GwgEAl2WJUuWvGdmg5vfMvdI2y+Z2QQf6bEE+JWZ3SPpfuBbhRA0nfLychYvXtz8hoFAILCdU19fz7Rp05gz509s2lRLaWmc008/jmuvvZbi4uj1fiT9p/mt/LY5Om3n4dIYlALvmdn0jBsWkMrKSgsKPxAIdCUSiQS33XYb11xzM6+99ir19UZxcT21tQkaGnoBV+PGri0DLmX48C2sWvVSZKUvaYmZVUbZNmuLZnaqpL2BOj/SMhAIBAJ5kEgkOOGEU1m06HVqay8C9sEp9p/gYmA+xg1v+gquq/Ro3n67kgsuuICZM2cWXJ6cA6/M7Pmkspc0TtIJLczVEQgEAt2KRCLB+eefz5///Dy1tY/RqNS/gktTVovLNv0qkEzR1Av4MbNn352pyVaTa+DV3Snzx+HylB8D/EnSGW0iTSAQCLSSRCJBTU0NO+44gR49BtKjx2DGjRtPTU0NicQ2iXXbTIYTTjiVG274I4nEZThFnkovXF7Auf5zVsq6vdm0qZa2IJeFn5qE6SJcBaVvAJ+hMX1rIBAIdAqSin7AgHJOO+1KVqz4EfX1T1Bffz0rV5bwjW9cxAknnNouSv+2225j0aLXSSTKcG6cTOyNK+mR/EzyPKWl8TaRK5fCT+3NLTazFQC+YET7PCYDgUAgAkmL+qyzLmXDhj64wMJUF8rjNDSM5IEHnmDBgpZkuM6Pa6652fvsx+J89pl4HlfuIfkJ8AlwKWeccXybyJVL4e8rVxj5I2CCL5CBXBHilhaNDgQCgYKTtKi3bNkB+DHZXCibN/fl6qtnbdtAgVm9ehXOsj8TuBKnyFP5xC8/DZf2/2hcdupKhg/fwq9//es2kSurwjezIjPrY2a9zaynLz4BrkLNuVEP4IsbPCvp3tYKGwgEuieJRIL58+dTWXk4Q4fuTGXl4cyfP3+re6bRon6T3C6U9axZ49wn9fX1VFdXE4+PJBbrRzw+kurqaurrW1xMbSujR4/BWfZVuLeMg3EK/VWSih3iFBVdRSz2H2AGZWXTqK4+NK+QzLxprugtrj7lRD8Nzadgrt//Qlxxh2aLiVdUVFggEAik0tDQYMcdV2Xx+P4Gdxq8anCnxeOVdvzxU6yhocGGDNnJLx/kt7G0qcFgmsFQKy4eYBMnTrK+fYca7NmkTdjThg/fxerq6lol87x58ywerzTY5I893+Awgx0NhlhpaR+bOHGSzZ8/3xoaGlp1LGCxRdXHWVfAfrjqPC/jquwswlVCegLYL1LjLjXvQ7iqQUHhBwKBvHHKc3+vPFOV+CaLxytt/vz5VlFxmFfYexjsk7ZtncGnDcanKfc9DD7j1ze2CXtadXV1q2RufEhV+mMt3+YhVSgKpfCX4ooupy//FPBcpMZd+bkKYFI2hQ+cg0vKtnjMmDEFuwiBQKBz09DQYPPmzbOKisNsyJCdrKLiMJs3b942yrBRmadb7WZw59b9nHKdZTDcYC+/z8sGu3tLftsHhlv+aW+FN7ZZVjaiIOeXfBgNHerOrxAWfTqFUviv5Vj3erMNu1qXv/XzWRV+6hQs/ECga9PQ0GBz58618vJ9TOprMMS7Wl7JagE3umsyKfzlNnToTlst6rKyCoNDDMYa7GRQ5o+R/YEBw7zLpbFNqW8HXqX8yEfh54rS+Yuk+yR9TdKn/fQ1SfeRuQhzOp8BjpW0EjeM7HCfnycQCHQjkh2uFRWHUVo6hNNO+xkrV16C2dPAb3G1y2cAX6a29h88+OCrTUInGztAM/E8o0aNIRaLcddd87jppguZODFGnz6bKCvbRCxWiqt5n6sjt5imA5/aLg6+w8n1NAC+AFwP/NlP1wNfjPo0SWlnEsHCDwS6HU07XKcZTMjiWqlMsbKdmyZJ0w7QpvslffjZcG8HBzVj4R/k3wYa3Tyt9eG3JxTIwsfM/mJm55nZMX46z8zub8sHUCAQ2P7IFjY5f/58nzjsMdwAox+SPc1A0sree2voJEBVVRVHHLEL8XjT0MZ4/GCOPHJXTj755KxyubeD/ckeC38pruzHENojDr7DifpkSJ2Ac1qyX3NTsPADge2PXGGT/fqNTbGuc/viG63sphZ+8hgt6QCdN2+e9+uf5N8iGiNmXIftZ8xF65RaWdkIq66ubnVIZntDHhZ+S6P7VbhHTiAQ2J5JjnJ1VnzSet+F2tqjgRE0+s+Tvvhd/PcEMB/4KbAWqAP2oWfPj/jOdy5vcoxYLMaUKVOYMmVKXrJVVVVxxx338eCDr7Jx42eBq4DV/lj9KS3dxFFH7cfChS8Qi+V0eHQJWnSGZnZDoQUJBALbJ42jXDO5asbilHwC2B1XLG8nYF+cG+VbwPvAKcBTwKXU1ZVx5533FSTJWWpnbkXFMvr0WUVZWYI+fYyJE4cxa9Z3Wbiwplsoe8idHlmSTpJ0op+fLOlaSdWSusfVCQS6IYlEgrlz5zJkyE5IfZH6U1LSn6lTp2ZMO9CYNyYTpwA/wqUYeAKYiQvyuwQYCRwGXEdqpI7ZEh544JWCJTlLvh0sXvwwGzasobb2TTZseJclSx5hypQp3UbZQ24LfyZwEvB1oAY4D3gaOAS4pu1FCwQC7U19fT37738Ip5/+M9at+zluTOQstmwp5/rr/8jo0Xtso/Rzh02OpbT0XeAFnFJPzWD5JLAC2Az8g8ZCIL3YvPmH7ZLkrLuRS+EfbGZfxf0yXwBOMbMa4FTcYzkQCHQhEokEBx54CM888wGZ0gvDaN55ZxMXXHBBk/2mTz+TeDxzFEw8/nPi8X446z2Ty+dS4Jc0F6kTKAy5FH49gJnVAU+b2Rb/vZ6QDz8Q6HLcdtttLF26guzK+WJg8Dbl95oLm1y/fj25Bz6tTplPKnk3oCpQWHIp/Hck7QBgZp9PLpQ0DNjS1oIFAoHC0Vx6YXCdr4lEMc2lF04vv5fsGL3xxulUVMxk6NAvUlExkxtvnM7ChTU4NZOrCEgsZX4M8AklJVdw4YVntfyEAxmRC+PMYwcpDsTNbG2hhamsrLTFixcXutlAoFuTrAblQicvwin0ZcTjV3LkkbtujVIZOnRn1q4dAnwX58ZJZyFwBWVla6mtfTPy8ceNG8/KlSU4t1Dqm8MnwEE4+3EJLhvLwRQVPcwxx+zFwoXzulWHakuRtMTMKqNs2+zVlNQj9buZ1RJcOoHAdkPTOPlGv3wyb82tt97K/Pnz+fjjTcC/gWrgFpr+zZMVmtblXX5vxozvU1T0LtsWATkYeBc4ApiI9G/Kyx9h9uz/F5R9W5FtRBauY3YN8B7wN6A8Zd0zUUd25TOFkbaBQHTq6ups6tSpVlY2wqS+VlY2wqZOnbrNSNHG9MINBvPMFeLYyWBfgzEGvbfJWulyxx9qLr3wnT4HzlAbNmznvEeiNjQ02LHHnmwlJeN8O+UGE0waaT17DixYIZDuCgVKj/w0sKef/yrwGvAp//3ZqAfIZwoKPxBonoaGBps9e7YVFw+wKBWbXAKxVwyqDPbfZnu3bIFPPTDFPxg2eaVfatDPpB3svPPOa3HagfbKDd8dKZTCfy7t+57AcuD4YOEHAu1HaqGQwYN3tL59R1ssNtBckY/MRT1Ssz06C3+aV+y5ioDUbpO1sqxsRFDMnZx8FH4uJ1mdj8hJun5eBCYDl9GYDCMQCLQhyQ7Xc8/9NUuWfIt166rZsGEQiUQJ7q+YKXzyx01CJ6dPP5NY7HYgW/qDHwNvAHeTHgvfu3dptxuN2pXJ9StejCtgvhUzWwMciuu9CQQCBSRT6OT555+f1uF6H/ADoJZc4ZOpoZNVVVX06FGfc/vGIiAhFr4rkzVbppktyrJ8A/A/bSZRINANyRQ6uXbtMp599nwSid/QaJkn89bEaZp5MpWmFZtisRh77bU3S5Zk3x5G+7YbY+FjsUu58ML/LtQpBjoBuZKn/VnSMelhmX7djpJmSPpm24oXCHQPsoVOJhJlNLXMk3lrjsOlJchc1CM9dNK5dbJtfyWuCMgo4Cf+eJVMnDggZ3GRwPZHLpfO2bhA2VckPS3pfkkPS3oDuAFYYma/bxcpA4EuTvYUw8n0wknOxCnon+MGLFXSNLY9c8Wmqqoq9tuvf4btD8alK34UeB1YibSAioq+PP74w8F339WI0rMLlOOGxE0AyqL2COc7hSidwPZOMqJmv/0OtdLSPiYNzBkjn9y+uHiQj08/zMfKN/hImXkGE1Oiaxp8eGWlD6WcYjDUoI8VFQ3MeIwkdXV1VlFxkMViwwxGmavleoHBHhaL9bfevQeHmPjtEAoRltkRU1D4ge2RpNKeOHGS9ew5yMevj/Zhk9vGyG/evHnr9sXF/Q0G+OlTPnxyX6/UG/x0qMViyXj75Qa3W0nJOOvXb6wNGZJfTHuIh+965KPw886l05aEXDqB7Y2mna0HAf8EJuLyxjzNtrljKigvh7ffLmHz5h+QzGsDlwP9gI+BnYEXcWGUX6Ws7LOcccaBPPnky6xZs4pRo8Zw4YVncfLJJweXSyCvXDotrWkbCHQrEokEt912G9dcczOrV69i9OgxTJ9+JolEIqWz9YvA94FpwLVkjnmfzMqVD+GKgTTWf4WjcR2n04HfkhzyEo//miOP3I3rrrsuKPdAq4mk8CWVAmPMbHkbyxMIdDqyhUyee+6V9Oixjtraq3DKOxkymStGfhHOms9WDORyXEnAq4B13HjjjGDJBwpGlGyZxwBLcYUokTRB0j1tLVgg0FnIlW1y/foPaVTuyZDJZIx8Jt6m+WIg7rO4uGcY5RooKFHupMtw75rrAcxsKTCuDWUKBDoV2UMme9E0bDIZMnk02WPkofliIM8DdYwePaSVkgcCTYmi8OvMja5NpfP09AYCbczq1UlXTSZOQboEp8yrcNb/EuBDMsXIx2INwBVkfhhcAQwCLkXazIwZ3y/0qQS6OVEU/ouSpgBFknaRdB2uxykQ2O5IJBLMmTOHwYNHIg1A6ktJyTCmTp1KfX19xn1Gj066ajIxlj59PvL1XP8IXIIbzLSZWOx1pHOQ9qesbBrV1Ydy883X5igG8g7wIbHYOxxzzOeZMmVKYU8+EGgubhMow+XOedpPVwAlUeM+85lCHH6g0KSnFu7de6RBv7QY+dsNRlpx8eCtce3z5s3bGps+b948i8crM6YWjscrraamJnJse7ZiIDDCoI+Vl+9hNTU1IS4+EBkKOfAKODHKskJMQeEHCklDQ4Mdd1yVxePJoh+/NBjo87//3o9qHekHPZX5AU+/NLjd4vFKO/74KdbQ0JDSTmXK4Kc7m2yTr1xh8FOgUOSj8JsdeCXpGTOb2NyyQhAGXgVaSqY4+QMO2J25cxf76JpewOG4AU27Ax/h0gxnGvi0G3AT8fih3HjjdKZMmUIikWDBggVcffWsMPgp0KnIZ+BVVoUv6Qu4kSQnAX9IWdUHGG9mB7RW0HSCwg+0hExx8rCMWCyZWvgrfsudgbf851NsOwo2deDTdKCEioqZLF78cHudSiCQN/ko/FymyVvAYtw/YUnKdA9wVAQhRkt6RNJLkl6UdEEUgQKBfMkvtXAJLmQy28CnX9NY9Wlv1qxZRSDQVciq8M3sOTObA+xsZnNSprvM7IMIbdcD3zWz8cCngG9JGl8guQOBreSXWjhBtIFPrhhIR1V8ylT9av78+SQSiQ6RJ9A1iOJ8PEDSg5JelfSGpBU+J35OzOxtM3vGz38EvAyMbKW8gcA2ZI+TPxNX0CMZ814FFBFt4NMoX/HprAJL2zzpdWzXrv0LS5Z8i3PP/RVf+crXg9IPtJgoCv9m4Grgs8D+uNEk++dzEEnlwH7AkxnWnSNpsaTF69aty6fZQBcjkUhQU1NDefkeFBX1RhpAUVF/xo3bl5qamqyKLnucfBXQm1gsOQDqdaA3zQ98ugJ4vcMqPuVK5fDgg6+yYMGCdpcp0EVoLowHeDJqyE+W/XfA+f5PaG7bEJbZ/UjNJd+jx0CDIQbDDPZrkku+qGgvO/74qoyhi7ni5MvKKqy6unprCOTYsbv7AiBNQyzd92E+Hr6/TZx4UNZCIq05z4qKwzLG+qdSUXGYl8kyTHdaRcVhBZMrsP1DgePwrwR+gat4NTE5RWocegB/BS6Msn1Q+N2LZHx7WVmlwecMBhn09Yr39ylVn5zyLimZYPPnz8/aTpQ4+Y4Y+LTteIBXc8bxDxmyk98mk8JfbkOH7lQw2QLbP4VW+I9kmB6OsJ+AucCvogoTFH7XJZOFe95551mPHuUGgw36+EFROxt826DCXPm+VKWf3brNZzBTew98cm8g+2cdqZv+EAsWfiAfCqrwWzrhfP6Gc64u9dMXc+0TFH7XJLOFe7u35sdb0zKA+xgMNzjJK/352711m68Cby6VQ6a3nED3JR+FH7UAytHAnqTEvZnZjGb6Bv7XW/mBbk7TTsjkLfQELmhrCdtWftofl7bpWFw8fDKJWMeFSbaG3Nk2t431r6qq4o477mPRooOprb0YFyb6PPH4lRx55K4d0pEc6BpEKYByPfA14Ns4BX4iLsA5EIhE5jj5q4AZZB4AdRnu1nwKFw8P8AklJVd0SJhka8mdbXPbh1gsFuOuu+Zx443TqaiYydChX6SiYiY33jidhQtrQiqHQMtp7hUAWJb2uQPwj6ivEPlMwaXTNcncCTkgZ8ck9DcYZXCQNRel09kJLppAW0IeLp0opsIm/7lR0gigDhhe+EdPoLNTX19PdXU18fhIYrF+xOMjqa6uzppHPklmCzeWYVmS5ACozUjPMW7cDGbPvoiFC+dtl9ZtVVUVRxyxi8+Z35gDPx4/OLhoAu1Lc08EXEXlfrgRIO/ginJeHvWJks8ULPzOR0NDg82dO9fGjNnNR9LsmdbJuqcNH75Lzpj1zBbuHj4cclur1y3vY+PG7bZdWvSZCCmRA20FhUyPnIqkEqCXbVvysCCEbJmdi0QiwZe/fAr33fd3Ghq24F7snmbbLJOVVFcfysyZM7O247JZvpbSCfkbXBLWMbhkZa5j0g37WEWvXlvYsOFdevbs2WbnFwh0BQqSHjmlsROBB8zsI0k/xA28utzMnm29qE0JCr9jSc8pX1bWkzVrNlJfPwR4E7iWxlTDqSykrGwatbVv5mw7NZ/8yJGjgXpefPFNNm/uC6zHvUiuo7y8D8uXLw3KPhCIQKEV/jIz20fSZ3FJRn4BXGJmB7Ze1KYEhd/+JLMyXnLJL1m5ciXOd14PDAW2AANwXr0zcdb9LhlaeRXpABKJ9XkfOxQVCQRaRz4KP0ocfoP/PBq40czuk3RFi6ULdBqSLpt7732BROIyGqs/zQDW4n761cBeQNyvy6Twn6e0NJ738WOxGFOmTAnFugOBdiKKGfWmpBtwsfj3ez9+ML+6APPnz/fK/mlSszK6pKaDcBG4pcDvgONwBUIyZZm8lDPOOL7d5G5vQm76QFchikunDPg88LyZvSZpOLC3mf2t0MIEl07ql2g+AAAYgUlEQVT7suOOE1ix4kdk88vDRbh0wu8Ar+G6b3oCP6axk/VShg/fwqpVL1FcHGng9nZFtvKJyVGvYSBUoKMpVIlDAMxsI/Bv4ChJ5wND2kLZBwpDfX09U6dOpaRkGFJfpAEMHjySOXPmbGORrl69mtzVn97HdaZuBA7GFfreF/gWsD9FRecydeohXVbZQ8hNH+haREmtcAEwHxjip3mSvt3WggWikepuGDx4R0pLh3L99f9gy5aZuJLEN/Hee/0544zpfPnLp6Qp/SiDn/oh9WTw4PcpLj6f4uIHGDduGDU1M9myZS2//e1vu6yyh1zlE3tRW3sxV189qyPECgRaRJR/6pnAgWZWCyDpZ8DjwHVtKVigeZy74RQeeGA5mzf/ALgLeA6n6NMTklVy//0PsWDBgq2dpKNGDWblyiv8+vTY+itw0TrrmDhxV5566h/d0nWRb+KzQKAzE+UfLBojdfDzIQtmJ8B1ur7I5s3/wrkbHsX51zMlJPsx9fWxJhbpjBnfJxZ7B+euaRzy776/DSQoL+/DE0/8vVsqe8g/8Vkg0JmJ8i++BXhS0mWSfozLa3tz24oViMKll15FQ8OlNCr4WnL75D9pYpGecsopfOlLhyK9g7Poj/KfbwEb2W+/PXnttWXNumxS3Up9+44iHh9J375Dqag4bLuPZpk+/Uzi8SvJFJ0Uj1+5XWbvDHRjouRfwIVnTMOlSN4vat6GfKeQSyc/iovTM06OyFloA4ZuU2yjoaHBampqbNy4fa24eIAVFw+ycePGRy7zl618H0w0ONTKyioylvHbXsinfGIg0BFQ6IpXXuFf4BV+pHq2LZm6s8Kvq6uzs88+26QdDPr5RGX97PDDD7fNmzdn3Ke4eFCagp/qk5tlSki2pxUXDy54Kt5c5ftcYfBbCpYCOJ9C4IUkJD4LdGYKqvCBS3AhG5fhHMTPAT+MeoB8pu6q8Ovq6nzO+CEG+6VZyuOtpGRQRqVfXp6ecbLOXE3YZNnA5VszWkJ/O/bYkwuupJor3wdufWvrsOZbCDwQ6C4UWuEvx2XITH4vBZZHPUA+U3dV+FOnTjUYaZDNUh5vRxxxxDb7zZ0714qKhntLOqngF5grCp58S+hvgwePtDlz5rSJUsxc3CQ5LTfYyQpRizbfQuCBQHchH4UfpdP2LZqGfZTgUicGIpBIJLjlllvo3bs/0gCkPsRiA5k8eTJbtmwBYM6cPwGDcSNbM0XYzOChh57Zpu1TTjmFo48+lJKS/6Ox0/VKSkp24LjjvkBDwweYvc/atWs47bTT8oq0iZpOoLkoFpf+uPXRLCEePhAoAM09EYC7cQp+Ni5iZw0u4Pta4NqoT5YoU1ey8Ddt2mR77LGHt7R7+88Sg695K3y8lZQMts2bN5vU16C8GUu5d8bjtIV/OR/3Sa7yfYX04Tf3JtHaN4hAYHuFPCz8KAOv/uinJI8W8HnTJfnkk08oKxuG2ShgFo1ZKC8F/oobGLWUzZsP5Oijj6a0NM7Gjf3IlY1S6pHxWG2RcbJpOoHGAVy1tUfz4IMHNxm8VVVVxR133MeiRQenFDd5HvgJ0Juyst9w5JG7tbqM3+jRY1i7Nvv1CfHwgUAEojwVcBmz9vJTj6hPk3yn7dHCr6urs3POOceKigZujayBWM5oGVegu9rgTpMGeB/+CG8RR/fhN0dLo1qa64jNFNqZfMvo02eklZWNsD59htjEiZMKFs0SCoEHApmhwJ22k4D/AH8HHgNWAIdEPUA+0/am8Ddv3mx9+45MiYpJRtb0byZyZahX8M5V0zRKZ4I1jbDJHqWTi9ZEtXRG90mIhw8EMlNohb8E2C3l+67AkqgHyGfqzAo/c5x8X4NSgw1pSrFPM/74vn5yFn6y/XPPPddisaZx+JMnT7ZNmzZttdQHD97Rysv3sHHj9s1ptbcmqiVfC7+9CPHwgcC2FFrhL4uyrBBTZ1X4zcXJwyCDzSlKcWBEC795V01TS/12g2O2kSGTldsapR3cJ4HA9kM+Cj9KnN5iSbMkTfLTTbhexy5JpjDKHj12YO3aT4CxQDJRWbI61BJc1uijU1qZgBuvlrk6lKsVu4mSknXcd999OeVp2oG6BVeMpKkMmXKztybLY1VVFUccsQvxeNOkavH4wRx55K6t7oANBAIdRHNPBFzc/YW4UMy7gOlASdQnSj5TR1v4DQ0N9qUvneTdKXumWNFDU3zr2az2ASnfN/vvmUa89jOI22GHHRbJL9/UUo9utbfWLRPcJ4HA9gEFdunEgaKU70VAWdQD5DO1t8JvaGiw3//+99arV/80v3wPgw9SFGS0OPmmkTX7+XYa4/ClEjvnnHOsrq4usoxNO1Cjd6YGt0wg0D3IR+FHcek8hEunkKQUWFSgF4x2Z8uWLUyePBlpAEVFffjmNy/mk0/KcfHyi3GZn3cByml0ycSBZJw8uMIg1cBIv/wQnLvlJZwL5CDgTYYNG0td3TrMPsTsAxKJT7jhhhvyqhDVdCRr9NzswS0TCAS2obknArA0yrIs+34el4vndeDi5rZvawt/06ZNVlzc37tWphmMslz5a2Av/z01Tv4jg138uvTO297+LSFu5557bl6WfDaaWurzLFusfiarPbhlAoGuDwV26fyTlJTIQAXweIT9inDFz3fEDdx6Dhifa5+2VPgNDQ02aNBQaxwQdVgEv3x/P1/n3SlDzEXg7JX1ITF48NCCKPpUuRvjz5NROk1j9UMseiDQfclH4UfxLXwHuEPSW7jShsOAr0XY7wDgdTN7A0DSAuA4nN+jTaivr2fatGnMmnUndXXrcO6WBqCYoqKPaWgQrhRvL2CVX5erQlS9ny8GlgKjcJfgMrIlOautnVbQot6xWIy77prHggULuPrq37F69X+Ix3sBl1Nb+zGjR4/hwgunc/LJJ3fbMoSBQCAazWomM3ta0u7Abn7RcjOri9D2SGB1yvc1wIHpG0k6BzgHYMyYludD2bJlC0OG7MiGDXFgM7AnLn2/y2PT0HCJF2c2LqRxDPABufLXuMvzqp+/hJKSnmzZsgWz7A+JTZtqW3wO2WiLfDmBQKD7EckkNLM6M3vBT1GUfWTM7EYzqzSzysGDB+e179q1a5GENICSkkFs2LARp6BH4zpg0+PlR+P6oBcAZwLvAZnrlbo4+vVAJdI5TJ48nA8/XENpaZxcHadufSAQCHQ+2tIH8CZOwyYZRQHz6K9du5ahQ3fGWfI34RT6TUB/nGWf2eUCZcCNQBXwaZxr5yBSI1mgAulNNm3aiNmHJBL/x6JFi+jZsyenn34cbvBU5kFVZ5xxfKFOMRAIBAqKnM+/DRqWkv6QyThF/zQwxcxezLZPZWWlLV4cbRCvJJyyX0xT5d4Hp/wzuWleBSpx/cmz/P7X4dL8lwAJQOy55ygWL15Mr17pDw3XTzB69Hjeeacn7sGSTAd8KcOHb2HVqpcK6sMPBAKBXEhaYmaVUbaNZOFL2kfSsZJOSE7N7WNm9cD5uATwLwO351L2+ZPNki8mdwWmYlzQ0Dm4fuVbKSpKcPPNV/kKUet54YUXMip7gOLiYlavfonq6kMpK5uGdABlZdOorj40KPtAINCpadbCl/R7XM/nizgTGMDM7JuFFiY/Cz+bJT8K6OvXpSrtT3ARpf9m0KCBxGK9GD16LBdeeFaIcAkEAtst+Vj4UczRT5nZ+FbK1AYkLfl0hf8qMByn3GfQ6HJJRulUsnHjJ3zuc7uxcGFNUPSBQKDbEEXbPS6pEyr8D8ickTIGDMWF+5+F89mfhRsDdjXwGBs3/u822SUDgUCgqxNF4c/FKf3lkpZJel5SNid5u/Huu+/iLPYK0iNsXArhdbiHwi3AfrgHw1m4U+5Fbe3FXH31rA6QPBAIBDqGKC6dm4Gv4/wiiWa2bTeGDBnC22+/yvDhw3GKvAEXffMh8C4wCKfkfwJ8N0MLuXPCBwKBQFcjioW/zszuMbMVZvaf5NTmkkVg2LBhNDQ0UFNzHX379sP57m8H3qfR2u8NZMoM2TS7ZCAQCHR1oij8ZyXdKqkqn7DM9iIWi3Hqqafy/vsrmT//h1RUzGTo0C9SUTGT6upJlJV9jEtdnMonxONXcuGFZ3WEyIFAINAhRAnLvCXD4g4Py4xCIpHghBNOZdGi16itvZhkxE48fiVHHrlriNIJBALbPQUNyzSzb7RepI6haabJmaxZs4pRo0J2yUAg0D2JYuHvCvwOGGpme0naBzjWzK4oqCAua+b/AJ2ifyCNQbhMa52RIFvLCLK1jCBby2hL2caaWaTMk1EU/t+B/wJuMLP9/LIXzGyvVovZ9DiLo76WtDdBtpYRZGsZQbaWEWRrnig+jTIzeyptWX3GLQOBQCDQaYmi8N+TtBNgAJK+CrzdplIFAoFAoOBEGXj1LVwC+d0lvQmsAE5tA1lubIM2C0WQrWUE2VpGkK1lBNmaIXI+fElxIGZmH7WtSIFAIBBoC3K6dCTtLmmypB3MrDap7CV9vn3ECwQCgUChyKrwJU0D/gR8G3hB0nEpq3/S2gNLGiDpQUmv+c/+Wbb7uaQXJb0s6Vq5UldtSh6yjZH0Ny/bS5LKO4tsfts+ktZI+k1byxVVNkkTJD3uf9Nlkr7WxjJ93if+e13SxRnWl0j6g1//ZHv8hnnIdqG/r5ZJekjS2M4iW8p2X5FkktolAiWKXJJO8tftRUm3todcUWTz+uIRSc/63/SL7SXbVsws44RLlraDny/H1RK8wH9/Ntt+USfg58DFfv5i4GcZtvk08E9cVrQi4HFgUmuPXQjZ/LpHgSP9/A64iKZOIZtf/2vgVuA3bS1XHr/prsAufn4ELgCgXxvJU4TLi70jrszZc8D4tG2qgev9/MnAH9rpWkWR7bDkPQVM7Uyy+e16A48BTwCVnUEuXIGMZ4H+/vuQznLNcH78qX5+PLCyPWRLnXK5dGJm9jGAma0EJgFfkHQ1UAgr+zhgjp+fA2Sq/m24slU9cUVne+BSYbY1zcrmawQUm9mDAGb2sZlt7AyyefkqcIUB/tYOMiVpVjYze9XMXvPzbwFrgUiDRlrAAcDrZvaGmW0BFngZs8l8JzC5Pd4io8hmZo+k3FNP4Mq5tQdRrhvA5cDP2LYoRUfKdTYw08w+ADCztZ1INsMV3QZXlu+tdpJtK7kU/ruSJiS/eOX/JdyIsb0LcOyhZpYM73wHp5yaYGaPA4/grMC3gb+a2csFOHarZcNZqusl3eVf0X4hqagzyCYpBlwFfK8d5EklynXbiqQDcA/zf7eRPCNxRROSrPHLMm5jrg7zBmBgG8mTr2ypnAn8pU0laqRZ2SRNBEab2X3tJFMkuXD/y10l/VPSE+3Y3xhFtsuAUyWtAe7HucvblVxhmaeRNsDK/yFOk3RDlMYlLQKGZVj1g7R2TdI24UKSdgb2oNGyeVDSwWb2jyjHb0vZcNfuYFx1lVXAH4AzcPUDOlq2auB+M1tTaGO1ALIl2xkO1ACnm1mnqbPQGZF0Kq5026EdLQtsNSiuxt3vnY1inFtnEk5vPCZpbzNb36FSOaqA2WZ2laSDgBpJe7Xn/Z9V4ZvZmhzr/hmlcTM7Its6Se9KGm5mb/s/f6ZXry8DTyRdS5L+AhwEtFrhF0C2NcBSM3vD73M38CkKoPALINtBwMGSqnF9Cz0lfWxmWTvf2lE25CrQ3wf8wMyeaK1MOXgTGJ3yfZRflmmbNZKKca/a/9eGMuUjG5KOwD1MDzWzze0gVxTZegN7AY96g2IYcI+kY82scOlu85cL3P/ySTOrA1ZIehX3AHi6DeWKKtuZwOfBeS8k9cJ5TNrL7RRppG1bcQ9wup8/HRcRlM4q4FBJxZJ64Cyc9nDpRJHtaaCfpKT/+XBcId0Ol83MTjGzMWZWjnPrzC2Esi+EbJJ6An/0Mt3ZxvI8DewiaZw/7slexlRSZf4q8LD5XrWOlk3SfsANuGSF7aYUmpPNzDaY2SAzK/f32BNexrZU9s3K5bkbZ90jaRDOxfNGG8sVVbZVwGQv2x64/sl17SBbI+3dS5yccH7Sh4DXgEXAAL+8EphljT3fN+CU/EvA1Z1FNv/9SGAZLqJpNtCzs8iWsv0ZtF+UTpTf9FSgDliaMk1oQ5m+iCt4/G/cGwXADJyCAvenuwN4HXgK2LE9rlVE2RbhghSS1+meziJb2raP0g5ROhGvmXDuppf8//LkznLNcJE5/8RF8CwFPtdesiWnyCNtA4FAILB9EyqABAKBQDchKPxAIBDoJgSFHwgEAt2EoPADgUCgmxAUfiAQCLQhkn4vaa2kFyJse4ikZyTVyxWbSi4vSNLBoPADkcjnpu1OSCqXNKUF+5VK+ntqOg6lZTeVVCbpPkmv+D/6lVnaGuizMH6stMyokr7mFcSLkn6Wsvzbkl6QdL+PG0fSZyVdk7LNYEkP5HtugW2YjR9wFYFVuFDq9CyfG4HTzGxP39avJPXLV5Cg8ANRmU30m7YJfgRrV6UcyFvhA98E7jKzhpRll+OyT6bySzPbHZfC4zOSvpChrU+AH5GWO0nSQOAXwGSvKIZJmuxXnwLsA/wLOMonjPuRlwEAM1sHvC3pMy04v4DHzB4D3k9dJmknSQ9IWiLpH5J299uuNLNlQCKtjYIkHQwKPxCJTDctZL9xJc2WdL2kJ3Fpk1P3KZL0S29hLpP0bb98slwiuuf9G0WJX75S0k8lLZW0WNJESX+V9G9J5/ltJkl6zFvEy/2xY35dlW/zhaSV62WY7Zc9L2m6X362pKclPSdpoaSylPO5VtK/JL2R8rp9JS6NxVJJ0327v/BtLJN0bpZLegopI5GVIbupmW00s0f8/BbgGTJkzDRXnOh/2TZr5Y7Aa15xgxvI9ZXkIXHZZ8twA+FOBf5iZum/8d1e1kBhuRH4tplV4B7Uv426o1qTdLC9R3qFafudcNbsC2nLHqIxv/2BuNQE4N4I7gWKMrQzFZeKuNh/H4Ab8boa2NUvmwt8x8+vpDGP+DW40c29cRbOu375JJzC2xE3QvtBXKqEEbjX5MG43FEP49I2VwAPpsjUz38OTFl2Be5PmTyfO3BG0nhcKtzkce9N2ecc4Id+vgRXR2Jc2vn3BN5J+R7DjVYdRZaR0UA/XIqArCOB0/cF+uNyy5T7c18I/Nmv+zoub/w8fy0fBnpkaHMk8HxH33vb+5T638Hlt9pE09HmL6dtPxv4aoZ2hgPLgU+1RI6u/KodaGMk7YArUnOHGrNylqRscoc1dVkkOQJXdKQewMzel7QvsMLMXvXbzAG+BfzKf0/mJUkW5vkI+EjS5hRf5lPWmMzuNuCzOOv1UfNWrqT5wCE418WOkq7DJXJLWtZ7SboCp2B3AP6aIvfd5jIbviQpW+rnzwH7pLwB9MUl71qRss0gIDV7Y87spt4ldhtwbfL8omBmH0iaisvkmsC5b3by62pw2UqRdAlwLa7exWm4B+93/bmuxT00A4UjBqw3swnNbpmCCpB0MCj8QGto7satLeCxkpkiEynzye/J+zg9T0jWvCFeGe4LHAWcB5yE86vPBo43s+cknYFPxJUmA2QvAiTcW8Ffs6wHZ931SvneXHbTG3GumV+RJ2b2Z+DPAJLOAZo8gCWNAA4wsxmS/o5LAvhDXJKvB72cm/I9biA7ZvahpBWSTjSzO3z/yT5m9ly2fVSgpIPBhx9oMWb2IS4F7YkAcuwbYdcHgXOTnbmSBuBeU8vlaiCAczn8PU+RDpDLVhgDvgb8Ly4h2qGSBslFxFQBf5fLpBgzs4U4BTfRt9Eb11HZg2i+64/8Pkn+Ckz1+yNpV0nx1B3MVWMqkkuPi+XIburfNvoC38nzWuD3H+I/++PeJGalbXI5cImfL8U9JBM43z64bJMhMqsV+LfNx4Hd5KKwzsTdW2dKeg54EV8dS9L+cgVSTgRukPSib+Yk3JvpGb6/aKlSClRFJVj4gUj4m3YSMMjfkJea2c24G/d3kn6I6wRcgMsGmItZOEWyTFIdcJOZ/UbSN3DuoWJcutnr8xTzaeA3wM64Sml/NLOEXEHpR3DW931m9if/YLol2bELfN9//gh4Epe29kmaKvNMLAMa/B93Nq6OcDnwjLfc1pG5DOXfcC6nRdkaljQKlwv/Fd8eOB/9LEnH4jJUXuK3XYkrn9dT0vG4TIwvAb9OeQjPSHGZJdMvY2bP+EW34lxmq2nsaD8M50YItBAzq8qyapuoNzN7mswd8/Nw/S2tImTLDHQJJE0CvmdmX+poWaIgVyJwupl9vaNlyYWkx4Dj/FtJYDsnuHQCgQ7AW9WPqH3qILcIueI+Vwdl33UIFn4gEAh0E4KFHwgEAt2EoPADgUCgmxAUfiAQCHQTgsIPBAKBbkJQ+IFAINBNCAo/EAgEugn/H845H1iWjdSEAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplot(1, 1, 1, aspect='equal')\n",
    "\n",
    "plt.scatter(x_acp[blues, 0], x_acp[blues, 1], c=\"blue\",\n",
    "            s=60, edgecolor='k')\n",
    "plt.title(\"Projection de l'ACP avec noyeaux \")\n",
    "plt.xlabel(\"1er composante (42.19%)\")\n",
    "plt.ylabel(\"2em coposante (23.05%)\")\n",
    "#bonne"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Essayer l'acp avec noyaux, au cas où les données ne sont pas linéairement séparables\n",
    "from sklearn.decomposition import KernelPCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "kacp = KernelPCA(kernel=\"rbf\", fit_inverse_transform=True, gamma=1)\n",
    "x_kacp = kpca.fit_transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '2em coposante ')"
      ]
     },
     "execution_count": 199,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.subplot(1, 1, 1, aspect='equal')\n",
    "\n",
    "plt.scatter(x_kacp[blues, 0], x_kacp[blues, 1], c=\"blue\",\n",
    "            s=60, edgecolor='k')\n",
    "plt.title(\"Projection de l'ACP avec noyeaux \")\n",
    "plt.xlabel(\"1er composante\")\n",
    "plt.ylabel(\"2em coposante \")\n",
    "#mal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [],
   "source": [
    "poly_acp = KernelPCA(kernel=\"poly\",degree=2, gamma=10)\n",
    "x_poly_acp = poly_acp.fit_transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 201,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '2em coposante ')"
      ]
     },
     "execution_count": 201,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.subplot(1, 1, 1, aspect='equal')\n",
    "\n",
    "plt.scatter(x_poly_acp[blues, 0], x_poly_acp[blues, 1], c=\"blue\",\n",
    "            s=60, edgecolor='k')\n",
    "plt.title(\"Projection de l'ACP avec noyeaux \")\n",
    "plt.xlabel(\"1er composante\")\n",
    "plt.ylabel(\"2em coposante \")\n",
    "#séparation linéaire"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [],
   "source": [
    "poly_acp1 = KernelPCA(kernel=\"poly\",degree=3, gamma=10)\n",
    "x_poly_acp1 = poly_acp1.fit_transform(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, '2em coposante ')"
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.subplot(1, 1, 1, aspect='equal')\n",
    "\n",
    "plt.scatter(x_poly_acp1[blues, 0], x_poly_acp1[blues, 1], c=\"blue\",\n",
    "            s=60, edgecolor='k')\n",
    "plt.title(\"Projection de l'ACP avec noyeaux \")\n",
    "plt.xlabel(\"1er composante\")\n",
    "plt.ylabel(\"2em coposante \")\n",
    "#séparation linéaire"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification non-supervisée"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {},
   "outputs": [],
   "source": [
    "#classification avec la méthode de K-means\n",
    "from sklearn.cluster import KMeans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 234,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '4 classes')"
      ]
     },
     "execution_count": 234,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "y_pred = KMeans(n_clusters=4, random_state=42).fit_predict(x_acp)\n",
    "\n",
    "plt.subplot(111)\n",
    "plt.scatter(x_acp[:, 0], x_acp[:, 1], c=y_pred)\n",
    "plt.title(\"4 classes\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_kmeans=kmeans.predict(x_acp)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 233,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '6 classes')"
      ]
     },
     "execution_count": 233,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "y_pred = KMeans(n_clusters=6, random_state=42).fit_predict(x_acp)\n",
    "\n",
    "plt.subplot(111)\n",
    "plt.scatter(x_acp[:, 0], x_acp[:, 1], c=y_pred)\n",
    "plt.title(\"6 classes\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 230,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '3 classes')"
      ]
     },
     "execution_count": 230,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "y_pred = KMeans(n_clusters=3, random_state=42).fit_predict(x_acp)\n",
    "\n",
    "plt.subplot(111)\n",
    "plt.scatter(x_acp[:, 0], x_acp[:, 1], c=y_pred)\n",
    "plt.title(\"3 classes\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '2 classes')"
      ]
     },
     "execution_count": 228,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "y_pred = KMeans(n_clusters=2, random_state=42).fit_predict(x_acp)\n",
    "\n",
    "plt.subplot(111)\n",
    "plt.scatter(x_acp[:, 0], x_acp[:, 1], c=y_pred)\n",
    "plt.title(\"2 classes\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Optimisation de k?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "En général, il y a pas un moyen de tester la performance d'une classification non-supervisée. Cependant, on peut optimiser le choix de k, le nombre de classes pour la méthode de k-means: on parle de x-means. La méthode fonctionne de la manière suivante: Par rapport à un certain critère heuristique, on choisit, disant la moitié des centres de classes pour une classification. On divise chacun de ces centres pour créer de nouveaux centres de classes. Si cette opération améliore la performance du modèle, on garde les et on itère jusqu'à convergence, sinon, on choisit d'autre ensemble de centres de classes.\n",
    "\n",
    "Pour mesurer la performance de chaque modèle, on utilise la métrique BIC dont l'expression est la suivante:\n",
    "\n",
    "$$ BIC(D | M) = L(M | D) - (p / 2)  log N $$\n",
    "\n",
    "$$D$$ la base de données \n",
    "$$M$$ le modèle pour chaque choix de k\n",
    "$$L(M | D)$$ la fonction de vraisemblance dépendante du modèle M étant donné une base D\n",
    "\n",
    "Nous avons installé la librairie pyclustering qui contient cette méthode implémentée."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 257,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-7.00725409e+11],\n",
       "       [-1.83239569e+10],\n",
       "       [-2.46069030e+10],\n",
       "       ...,\n",
       "       [ 6.79463843e+02],\n",
       "       [-2.13162165e+02],\n",
       "       [ 1.46260483e+02]])"
      ]
     },
     "execution_count": 257,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "A = np.asarray(x_acp).reshape(-1,1) #convertir la matrice x en une liste de 1-D\n",
    "A"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 258,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-258-2c224656b31c>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[0mvisualizer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend_clusters\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclusters\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mA\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     19\u001b[0m \u001b[0mvisualizer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend_cluster\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcenters\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmarker\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'*'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 20\u001b[0;31m \u001b[0mvisualizer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/.local/lib/python3.6/site-packages/pyclustering/cluster/__init__.py\u001b[0m in \u001b[0;36mshow\u001b[0;34m(self, figure, invisible_axis, visible_grid, display, shift)\u001b[0m\n\u001b[1;32m    593\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    594\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mcluster_descr\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcanvas_data\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 595\u001b[0;31m                 \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__draw_canvas_cluster\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdimension\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcluster_descr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    596\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    597\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0mattribute_descr\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcluster_descr\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mattributes\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/.local/lib/python3.6/site-packages/pyclustering/cluster/__init__.py\u001b[0m in \u001b[0;36m__draw_canvas_cluster\u001b[0;34m(self, ax, dimension, cluster_descr)\u001b[0m\n\u001b[1;32m    639\u001b[0m                     \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcolor\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcolor\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmarker\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmarker\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmarkersize\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmarkersize\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    640\u001b[0m                 \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 641\u001b[0;31m                     \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcolor\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcolor\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmarker\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmarker\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmarkersize\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmarkersize\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    642\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    643\u001b[0m             \u001b[0;32melif\u001b[0m \u001b[0mdimension\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.6/site-packages/matplotlib/__init__.py\u001b[0m in \u001b[0;36minner\u001b[0;34m(ax, data, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1808\u001b[0m                         \u001b[0;34m\"the Matplotlib list!)\"\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlabel_namer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1809\u001b[0m                         RuntimeWarning, stacklevel=2)\n\u001b[0;32m-> 1810\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1811\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1812\u001b[0m         inner.__doc__ = _add_data_doc(inner.__doc__,\n",
      "\u001b[0;32m/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_axes.py\u001b[0m in \u001b[0;36mplot\u001b[0;34m(self, scalex, scaley, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1613\u001b[0m             \u001b[0mlines\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mline\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1614\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1615\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautoscale_view\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mscalex\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mscalex\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscaley\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mscaley\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1616\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mlines\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1617\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_base.py\u001b[0m in \u001b[0;36mautoscale_view\u001b[0;34m(self, tight, scalex, scaley)\u001b[0m\n\u001b[1;32m   2426\u001b[0m             \u001b[0mstickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0martist\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msticky_edges\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0martist\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_children\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2427\u001b[0m             \u001b[0mx_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msticky\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mx\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0msticky\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstickies\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2428\u001b[0;31m             \u001b[0my_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msticky\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0my\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0msticky\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstickies\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2429\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_xscale\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'log'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2430\u001b[0m                 \u001b[0mx_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mxs\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mxs\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mx_stickies\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mxs\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_base.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m   2426\u001b[0m             \u001b[0mstickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0martist\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msticky_edges\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0martist\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_children\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2427\u001b[0m             \u001b[0mx_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msticky\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mx\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0msticky\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstickies\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2428\u001b[0;31m             \u001b[0my_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msticky\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0my\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0msticky\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstickies\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2429\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_xscale\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'log'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2430\u001b[0m                 \u001b[0mx_stickies\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mxs\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mxs\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mx_stickies\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mxs\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "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": [
    "#trop long: nous avons laissé l'éxecution pour 40 minutes sans avoir de résultats\n",
    "from pyclustering.cluster import cluster_visualizer\n",
    "from pyclustering.cluster.xmeans import xmeans\n",
    "from pyclustering.cluster.center_initializer import kmeans_plusplus_initializer\n",
    "\n",
    "\n",
    "# Préparer les centres initiaux - le nombre de centres initiaux définissent le nombre de classes par lesquelles\n",
    "# on commence l'analyse\n",
    "\n",
    "nombre_centres_initiaux = 2\n",
    "centres_initiaux = kmeans_plusplus_initializer(A, nombre_centres_initiaux).initialize()\n",
    "# création d'une instance de l'algorithme de x-means. l'algorithme va commencer l'analyse avec 2 classes, le nombre maximal\n",
    "# de classe est 10\n",
    "xmeans_instance = xmeans(A, centres_initiaux, 10) \n",
    "xmeans_instance.process()\n",
    "# Extract clustering results: clusters and their centers\n",
    "classes = xmeans_instance.get_clusters()\n",
    "centres = xmeans_instance.get_centers()\n",
    "visualizer = cluster_visualizer()\n",
    "visualizer.append_clusters(classes, A)\n",
    "visualizer.append_cluster(centres, None, marker='*')\n",
    "visualizer.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 325,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Clustering par la méthode AgglomerativeClustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 326,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 327,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 2, 2, ..., 1, 1, 1])"
      ]
     },
     "execution_count": 327,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clustering = AgglomerativeClustering(n_clusters=5).fit_predict(x_acp)\n",
    "clustering "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 328,
   "metadata": {},
   "outputs": [],
   "source": [
    "#visualisation avec la méthode AgglomerativeClustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 329,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, '5 classes')"
      ]
     },
     "execution_count": 329,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "plt.subplot(111)\n",
    "plt.scatter(x_acp[:, 0], x_acp[:, 1], c=clustering)\n",
    "plt.title(\"5 classes\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On préfère utiliser les k-means puisque sa compléxité est plus petite en temps d'exécution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 252,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#dendogramme\n",
    "from scipy.cluster.hierarchy import dendrogram, linkage  \n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "linked = linkage(x_acp, 'single')\n",
    "\n",
    "#labelList = range(1, 7200)\n",
    "\n",
    "plt.figure(figsize=(10, 7))  \n",
    "dendrogram(linked,  \n",
    "            orientation='top',\n",
    "            distance_sort='descending',\n",
    "            show_leaf_counts=True)\n",
    "plt.show()  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "On peut considérer des classes pour la classification mais c'est déjà fait avec l'algorithme des K-means."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Considèrons maintenant qu'on connait la classe pour chaque observation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Il s'agit d'apprentissage supervisé"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 262,
   "metadata": {},
   "outputs": [],
   "source": [
    "#séparation de la base de données en base d'apprentissage et base de test\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 263,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.20, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 264,
   "metadata": {},
   "outputs": [],
   "source": [
    "#classification avac support vector machine\n",
    "from sklearn.svm import SVC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 266,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SVC(C=1.0, cache_size=200, class_weight=None, coef0=0.0,\n",
       "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
       "  max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
       "  tol=0.001, verbose=False)"
      ]
     },
     "execution_count": 266,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_svc=SVC(gamma='auto') #kernel rbf par défaut\n",
    "clf_svc.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 267,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.40160642570281124"
      ]
     },
     "execution_count": 267,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#mesurer la performance de cette classification\n",
    "clf_svc.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Mauvaise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9879518072289156"
      ]
     },
     "execution_count": 270,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_svc=SVC(gamma='auto',kernel='linear')\n",
    "clf_svc.fit(X_train,y_train)\n",
    "clf_svc.score(X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "metadata": {},
   "outputs": [],
   "source": [
    "#très bonne modèle: résultat attendu car les données sont linéaire d'après la visualisation de l'ACP"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "metadata": {},
   "outputs": [],
   "source": [
    "#classification par arbre de décision\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,\n",
       "            max_features=None, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, presort=False, random_state=42,\n",
       "            splitter='best')"
      ]
     },
     "execution_count": 279,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_decision_tree = DecisionTreeClassifier(random_state=42)\n",
    "clf_decision_tree.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.92592593, 0.92307692, 0.96153846, 1.        , 0.92      ,\n",
       "       1.        , 0.95833333, 1.        , 0.83333333, 0.95652174])"
      ]
     },
     "execution_count": 282,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "cross_val_score(clf_decision_tree, X_test, y_test, cv=10) #appliquer la cross validation pour arriver à la combinaison \n",
    "#optimale des hyperparamètres du modèle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "metadata": {},
   "outputs": [],
   "source": [
    "#bon précision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "            max_depth=2, max_features='auto', max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, n_estimators=100, n_jobs=None,\n",
       "            oob_score=False, random_state=42, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 288,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#classification avec les forets aléatoires\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "clf_foret_alea=RandomForestClassifier(n_estimators = 100, max_depth=2, random_state=42)\n",
    "clf_foret_alea.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7710843373493976"
      ]
     },
     "execution_count": 291,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf_foret_alea.score(X_test, y_test) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 292,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.85185185, 0.80769231, 0.84615385, 0.76923077, 0.88      ,\n",
       "       0.95833333, 0.79166667, 0.875     , 0.58333333, 0.82608696])"
      ]
     },
     "execution_count": 292,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(clf_foret_alea, X_test, y_test, cv=10) #améliorer la performance du modèle par la cross_validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "metadata": {},
   "outputs": [],
   "source": [
    "#bon modèle si on choisit les bons hyperparamètres"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 309,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Classification avec la régression logistique\n",
    "from sklearn.linear_model import LogisticRegression \n",
    "clf_lin_reg=LogisticRegression(random_state=42, solver='lbfgs',multi_class='multinomial', max_iter=1000).fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1.        , 1.        , 1.        , 1.        , 1.        ,\n",
       "       1.        , 1.        , 1.        , 0.91666667, 0.95652174])"
      ]
     },
     "execution_count": 311,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cross_val_score(clf_lin_reg, X_test, y_test, cv=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#On remarque qu'il y a des modèles parfaits: score 1, donc soit il y avait du sur-apprentissage où bien il y avait une fuite\n",
    "#de données de test lors de l'apprentissage!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Notre modèle final qui performe le mieux est le SVC (support vector classifier) avec une précision de 98.79%"
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}