Notebooks/ch9_support_vector_machines_labs.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 9. Support Vector Machines – Labs\n",
    "\n",
    "Lab excercises from **Chapter 9** of [An Introduction to Statistical Learning](http://www-bcf.usc.edu/~gareth/ISL/) by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "metadata": {},
   "outputs": [],
   "source": [
    "import itertools\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import patsy as pt\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set()\n",
    "\n",
    "from sklearn import preprocessing\n",
    "from sklearn import svm\n",
    "from sklearn import metrics\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.model_selection import cross_val_score \n",
    "from sklearn.model_selection import GridSearchCV\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9.6.1 Support Vector Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate som sample data with 2 predictors and a response\n",
    "np.random.seed(2)\n",
    "X = np.random.normal(0, 1, (20, 2))\n",
    "y = np.concatenate((np.repeat(0, 10), np.repeat(1, 10)))\n",
    "X[y==1, :] = X[y==1, :] +1\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>x1</th>\n",
       "      <th>x2</th>\n",
       "      <th>y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.416758</td>\n",
       "      <td>-0.056267</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-2.136196</td>\n",
       "      <td>1.640271</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-1.793436</td>\n",
       "      <td>-0.841747</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.502881</td>\n",
       "      <td>-1.245288</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-1.057952</td>\n",
       "      <td>-0.909008</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.551454</td>\n",
       "      <td>2.292208</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.041539</td>\n",
       "      <td>-1.117925</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.539058</td>\n",
       "      <td>-0.596160</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>-0.019130</td>\n",
       "      <td>1.175001</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>-0.747871</td>\n",
       "      <td>0.009025</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.121892</td>\n",
       "      <td>0.843566</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1.256570</td>\n",
       "      <td>0.011221</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.661178</td>\n",
       "      <td>0.763816</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.362345</td>\n",
       "      <td>-0.187612</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>-0.421217</td>\n",
       "      <td>0.846505</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.730943</td>\n",
       "      <td>3.231367</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>-1.434768</td>\n",
       "      <td>1.112727</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>1.370445</td>\n",
       "      <td>2.359634</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>1.501857</td>\n",
       "      <td>0.155786</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>1.000010</td>\n",
       "      <td>1.542353</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          x1        x2  y\n",
       "0  -0.416758 -0.056267  0\n",
       "1  -2.136196  1.640271  0\n",
       "2  -1.793436 -0.841747  0\n",
       "3   0.502881 -1.245288  0\n",
       "4  -1.057952 -0.909008  0\n",
       "5   0.551454  2.292208  0\n",
       "6   0.041539 -1.117925  0\n",
       "7   0.539058 -0.596160  0\n",
       "8  -0.019130  1.175001  0\n",
       "9  -0.747871  0.009025  0\n",
       "10  0.121892  0.843566  1\n",
       "11  1.256570  0.011221  1\n",
       "12  0.661178  0.763816  1\n",
       "13  0.362345 -0.187612  1\n",
       "14 -0.421217  0.846505  1\n",
       "15  0.730943  3.231367  1\n",
       "16 -1.434768  1.112727  1\n",
       "17  1.370445  2.359634  1\n",
       "18  1.501857  0.155786  1\n",
       "19  1.000010  1.542353  1"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot it, is it linearly seperable?\n",
    "df = pd.concat([pd.DataFrame(data=X, columns=['x1', 'x2']), pd.Series(y, name='y')], axis=1)\n",
    "display(df)\n",
    "\n",
    "plt.figure(figsize=(10, 10))\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df);\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = svm.SVC(kernel='linear', C=10, random_state=0).fit(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Decistion boundary plot\n",
    "\n",
    "x1 = np.arange(-4, 4, .1)\n",
    "x2 = np.arange(-4, 4, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df)\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10caba400>"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Contour plot\n",
    "\n",
    "x1 = np.arange(-4, 4, .1)\n",
    "x2 = np.arange(-4, 4, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.decision_function(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df);\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='black', marker='+', s=500)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### We can obtain some basic information about the support vector classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=10, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[6 6]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[-10.          -5.05497878  -7.40089843 -10.         -10.\n",
      "  -10.          10.          10.           2.45587721  10.\n",
      "   10.          10.        ]]\n",
      "\n",
      "Weights assigned to the features (coefficients in the primal problem).\n",
      "[[1.00241093 0.91678573]]\n"
     ]
    }
   ],
   "source": [
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nWeights assigned to the features (coefficients in the primal problem).')\n",
    "print(model.coef_)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### What if we instead use a smaller value of the cost parameter?\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=0.1, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[8 8]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[-0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1 -0.1  0.1  0.1  0.1  0.1  0.1  0.1\n",
      "   0.1  0.1]]\n",
      "\n",
      "Weights assigned to the features (coefficients in the primal problem).\n",
      "[[0.473289   0.29874954]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = svm.SVC(kernel='linear', C=0.1, random_state=0).fit(X, y)\n",
    "\n",
    "# Decistion boundary plot\n",
    "\n",
    "x1 = np.arange(-4, 4, .1)\n",
    "x2 = np.arange(-4, 4, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df)\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500);\n",
    "\n",
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nWeights assigned to the features (coefficients in the primal problem).')\n",
    "print(model.coef_)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With lower cost parameter the model uses more support vectors because the margin is now wider"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Use cross-validation to tune Cost parameter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>C</th>\n",
       "      <th>accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.001</td>\n",
       "      <td>0.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.010</td>\n",
       "      <td>0.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.100</td>\n",
       "      <td>0.60</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000</td>\n",
       "      <td>0.70</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5.000</td>\n",
       "      <td>0.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>10.000</td>\n",
       "      <td>0.75</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>100.000</td>\n",
       "      <td>0.75</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         C  accuracy\n",
       "0    0.001      0.60\n",
       "1    0.010      0.60\n",
       "2    0.100      0.60\n",
       "3    1.000      0.70\n",
       "4    5.000      0.75\n",
       "5   10.000      0.75\n",
       "6  100.000      0.75"
      ]
     },
     "execution_count": 215,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Cs = [0.001, 0.01, 0.1, 1, 5, 10, 100]\n",
    "scores = []\n",
    "for C in Cs:\n",
    "    model = svm.SVC(kernel='linear', C=C, random_state=0)\n",
    "    score = cross_val_score(model, X, y, cv=5)\n",
    "    scores += [score]\n",
    "    \n",
    "scores_mean = np.mean(np.asarray(scores), axis=1)\n",
    "\n",
    "pd.DataFrame({'C': Cs, 'accuracy': scores_mean})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,0,'log(C)')"
      ]
     },
     "execution_count": 216,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "sns.lineplot(x=np.log(Cs), y=scores_mean)\n",
    "plt.xlabel('log(C)')\n",
    "# plt.ylabel('accuracy');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate simulated test set\n",
    "xtest = np.random.normal(0, 1, (20, 2))\n",
    "ytest = np.random.choice([0, 1], size=20, replace=True)\n",
    "xtest[ytest==1, :] = xtest[ytest==1, :] +1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 6,  4],\n",
       "       [ 0, 10]])"
      ]
     },
     "execution_count": 218,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# With cost = 5\n",
    "# Test model selected by cross-validation\n",
    "model = svm.SVC(kernel='linear', C=5, random_state=0).fit(X, y)\n",
    "ypred = model.predict(xtest)\n",
    "\n",
    "confusion_matrix(ytest, ypred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[7, 3],\n",
       "       [2, 8]])"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# With cost = 0.01\n",
    "# Test model selected by cross-validation\n",
    "model = svm.SVC(kernel='linear', C=0.01, random_state=0).fit(X, y)\n",
    "ypred = model.predict(xtest)\n",
    "\n",
    "confusion_matrix(ytest, ypred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Now consider a situation in which two classes are linearly seperable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 231,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Generate set that is linearly seperable\n",
    "np.random.seed(1)\n",
    "X = np.random.normal(0, 1, (20, 2))\n",
    "y = np.concatenate((np.repeat(0, 10), np.repeat(1, 10)))\n",
    "X[y==1, :] = X[y==1, :] + 1.3\n",
    "\n",
    "df = pd.concat([pd.DataFrame(data=X, columns=['x1', 'x2']), pd.Series(y, name='y')], axis=1)\n",
    "plt.figure(figsize=(10, 10))\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=100000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[2 1]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[ -9.19595849 -25.03717651  34.233135  ]]\n",
      "\n",
      "Weights assigned to the features (coefficients in the primal problem).\n",
      "[[4.98467412 6.60267588]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmYAAAJVCAYAAAB0wjmJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3Xl8VOXd///3mWwkJCFsAcQFWVygGlcKakREQERAxa3aCgVvrT/bqrVavLU/2rrUpVVrrd62tWqr4k1VRG0RaG4rsUSjqLhRRSKVQIAQErIQss31/SNkJBvJDJlzrpnzej4ePOqc2T6dx1Hffj7nXJdjjDECAACA5wJeFwAAAIBmBDMAAABLEMwAAAAsQTADAACwBMEMAADAEgQzAAAASxDMAAAALEEwAwAAsATBDAAAwBIEMwAAAEsQzAAAACxBMAMAALAEwQwAAMASiV4XEK669QVSwx6vywAAAOhcUi+ljBof9ttiLpipYY9MQ63XVQAAAHTKifB9jDIBAAAsQTADAACwROyNMgEAQNxqNI5KTF/VKcnrUrotRQ0a4pQr0TEH/FkEMwAAYI0S01dJ6f3VNzVNjhPplVruMcaouna3SqqlQ5ydB/x5jDIBAIA16pSk9BgJZZLkOI7SU9N6rMNHMAMAAFaJlVDWoifrJZgBAABYgmAGAABgCYIZAACAJQhmAAAAHfj5XXdp6auvhh5fe8P1+mTdp1H9ToIZAABAB86dNk2vrVwpSSrZulUVFRUac/ToqH4nwQwAAKADJxx3nHaU7VDJ1q16beUKnT1latS/k2AGAADQAcdxdM7UqVr5f3n6x+uv6+zJk6P+nQQzAACATpwz9WwtefkVDc4epIEDBkT9+whmAAAAnRiUna3Bg7J1ztnRH2NK7JUJAADQIWOMdpSVqWznTuWecqor30nHDAAAoAOvr1qlOf91pb535X8pOTnZle+kYwYAANCBMydM0JkTJrj6nXTMAAAALEEwAwAAsATBDAAAwBIEMwAAAEsQzAAAACzBXZkAAAD7sSLvH3ry6afV2NioS2bP1uzzzo/adxHMAABATFu9bodeeHOzyqrq1T8jWbNPG6pTju6Z7ZNKS0v12OOP64n/eUxJSUm6+gff1wnHHa/Dhw3rkc9vi1EmAACIWavX7dCTK/+jsqp6SVJZVb2eXPkfrV63o0c+/5331ujE449XZmamUlNTNXHCBL2+6o0e+eyOEMwAAEDMeuHNzapvDLY6Vt8Y1Atvbu6Rz99RVqb+/fqHHvfv11/bS0t75LM7QjADAAAxq6VT1t3j4QoGjRzHCT02xigQiF58IpgBAICY1T+j4z0sOzseruyBA1S2syz0eGf5Tg3o338/7zgwBDMAABCzZp82VMmJreNMcmJAs08b2iOff/IJJ+rd995TeUWF9uzZo3+uWqVxJ4/tkc/uCHdlAgCAmNVy92W07socOHCgrp4/X9//0Q1qbGzUjHPO0eijj+6Rz+6IY4wxUfv0KKj79HWZhlqvywAAAFHwRXCQhmQP8bqMsJVsL9HIwLbQYycpVSmjJ4b9OYwyAQAALEEwAwAAsATBDAAAwBIEMwAAAEsQzAAAACxBMAMAALAEwQwAAGA/ampqdPm876pk69aofxcLzMIaTkpvOVlD5SSmSMFGBXdtldld7nVZAAAf+2Tdp7r717/WpuJiV76PYAYrOMlpcvofru1LH1L9ti+VkDlA2TN/oIATkKkp6/oDAAC+Vb/hHe157xWZmnI5vfuq1wkzlDzi5B757Jf/9jfd+MPr9Itf3tUjn9cVRpmwgtNniEr//j+q3/alJKmpcoe2PX+vAlkHeVwZAMBm9RveUe3qRTI1zRMWU1Ou2tWLVL/hnR75/Ft+fJOOO/bYHvms7iCYwQpOSm/VFX/W6lhwT41MU6NHFQEAYsGe916RmhpaH2xqaD4egwhmsIKp363kISNaHXOSU+UkJHlUEQAgFrR0yrp73HYEM1jB7CpR9vRrlJg1SJIUSE3XoPOuV7CyxOPKAAA2c3r3Deu47bj4H1YwdTUyFcUafOFNUiAgR46ClVtlqnZ4XRoAwGK9Tpih2tWLWo8zE5LU64QZ3hV1AAhmsMeeKgW3rvO6CgBADGm5+zJad2W2eHHRcz36eZ0hmAEAgJiWPOLkHg9iXuEaMwAAAEsQzAAAACxBMAMAAFYxxnhdQlh6sl6CGQAAsEaKGlRduztmwpkxRtW1u5Wihq5f3A1c/A8AAKwxxClXSbW0tbrS61K6LUUNGuL0zIK2BDMAAGCNRMfoEGen12V4hlEmAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAAACWIJgBAABYgmAGAABgCYIZAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAAACWIJgBAABYgmAGAABgCYIZAACAJQhmAAAg6pzUTAUyBkoJSV6XYjWCGQAAiJ6EZCUMPUYNDUY1xUVy+g9XoO/BXldlrUSvCwAAAPErMHC4tv/tf1RX/G9JUnn+XzXowpuU2LufTM1Oj6uzDx0zAAAQHYEEmWAwFMqaGZWv+l85af08K8tmBDMAABA1pqmp/bHGeslxPKjGfp4Es4cffljTp0/X9OnTde+993pRAgAAiLZgkxJSeimp30GtDvcZe65MbYVHRdnN9WvMVq9erTfffFNLliyR4zi68sortXLlSk2ePNntUgAAQJQFd3ypwZfcouqP81W/o1jp3zhdSRl9FNz+hdelWcn1YDZw4EAtWLBAycnJkqQRI0Zoy5YtbpcBAABcYOp3q6n4Q6UdMlK9hx0tU7tLwe2lXpdlLdeD2ahRo0J/vXHjRi1btkyLFi1yuwwAAOAWE5Sp3CbjdR0xwLOL/9evX6958+bp5ptv1rBhw7wqAwAAwBqeBLM1a9Zo7ty5uvHGG3X++ed7UQIAAIB1XB9llpSU6Nprr9UDDzyg8ePHu/31AAAA1nI9mD3++OOqq6vT3XffHTp26aWX6lvf+pbbpQAAAFjFMcbE1LV4dZ++LtNQ63UZAAAAnXKSUpUyemLY72PlfwAAAEsQzAAAACxBMAMAALAEwQwAAMASBDMAAABLEMwAAAAsQTADAACwBMEMAADAEq6v/A8gjiWmKDBgmBRIkuMEZOp3K1j2pRRs8royAIgJBDMAPcRRwpCjtf3lh1W3+TNJUtqR31T/CReracunHtcGALGBUSaAHuFkDFDNZ2+HQpkk7f7sbdXv3CqnV7qHlcFPnKRUOSm9vS4DiBjBDECPcBJTVFdS1O543dYv5SSlelARfCWplxKGHqOmlD5qMIlKODhHTmofr6sCwsYoE0CPMHuq1PvIsapZt7rV8bQRxyu4a7NHVcEvEgYdoa0v/FoNpV9JkgKpGRp6xR1qKvlUCjZ6XB3QfXTMAPQIU7tLKQOGKuu0CxVISVNC7ywNOPsqBRwjNdZ5XR7imNMrQ3VbvwyFMkkK1lZp17vLFMjM9rAyIHx0zAD0mKatn6n3sKOVcfQpkoIyNWUK7vjS67IQ55yEJDVVbG93vLG6XHISPKgIiBwdMwA9yMhUbFFTySdqKlmnYGX7f1kCPS24u0KpI46XElr3GjJPmCyzu9yjqoDI0DEDAMQ2E5Sp3Kah37ld5W8+r2D9HvUZO12JySkKVpV4XR0QFoIZACDmmartcuqq1e+UmZITkKktV7Bso9dlAWEjmAEA4oKp3y2z86uuXwhYjGvMAAAALEEwAwAAsATBDAAAwBIEMwAAAEsQzAAAACxBMAMAALAEwQwAAMASBDMAAABLEMwAAAAsQTADAACwBMEMAADAEgQzAAAASxDMAAAALEEwAwAAsATBDAAAwBIEMwAAAEsQzAAAACxBMAMAALAEwQx2SEyWk5opOZySAAD/4t+C8JYTUGDQEVLGENXv3qPAQWPkZA7xuioAADyR6HUB8LfAgMNV8fbfVP3xquYDTkCDL16ghN59ZWrKvS0OAACX0TGDt5JSvw5lkmSCKsv7i5zeA7yrCQAAjxDM4B3HkWlsaHc4WFspBWjmAgD8h2AG7xijQGKiErMGtTqcnnOmzJ5dHhUFAIB3aEvAU8Gy/2jIpbeq4q2XVb+jWOlHj1fa4ceqqeQTr0sDAMB1BDN4ytRVq2nLJ+rzjVOkhCSZ+ho1bfnY67IAAPAEwQzeCzYqWLHZ6yoAAPAc15gBAABYgmAGAABgCYIZAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAAACWIJgBAABYgmAGAABgCYIZAACAJQhmAAAAlkj0ugDAl5yAAun9JDkKVpdJJuh1RQAAC9AxA1zmpPZRwtBjtHvbZtVs2aiEg74hJ72/12UBACxAxwxwkxNQYMAwbX7yvxWsrZYkVax+SUPn3S2ze5cUbPS4QACAl+iYAS4KpPdTzb/fDoUySTINe1T1/j8UyBzoYWUAABvEXMcs+P4qmaqdYb0nMH5qlKoBwuXIBJvaHQ02NUpy3C8HAGAVX3TMggXLvS4BkCQFq8uUftR4OUm9vj4YSFSfEyYrWFXqXWEAACs4xhjjdRHh2PX/Xapg6daw3pMyKUcSnTPYwendX06fg1T5/kqZpkb1OWGKTM0OmartXpcGAOghTlKqUkZPDPt9MTfK/M97maov3hPWe0ZorSQpZe9jAhq8ZGrKZGp3KfPI4yU5CpZ+ITXVe10WAMACvhhlbijIkiTV5TUHNEab8FywUcGKEgUrthDKAAAhMdcxeyeQpspAeljvyQ1Wa0NBlkaMr1Bd3lqlTMpRsGA5nTMAAGCVmAtmhU61Sp3K8N4UyGz+34Lm/2G0CQAAbOSLUWb+3iCXv7fTxmgTAADYKObuyhyXM0XFm7aE9Z6xGcMlSbmmuXOWG2xe3HPE+ApJX9+1KdE9AwAAB843d2VGorCqqPkv9ga0TkebXHsGAAA85ItRZouWgMZoEwAA2MgXHbN9FVYVaWzGcOU7lco1mcoPpIfu2pS4MQAAAHjHVx2zFoVVRSqsKlK+U9n8J5BO9wwAAHjOl8GsBaNNAABgE9+NMtvqaLQpiTXPAACA63zdMWvRdrQp0T0DAADuI5jtg9EmAADwku9HmW3tO9qUxHZOAADANXTMOtAy2pTongEAAPcQzPaDcAYAANzEKLML3d7Oae/rGW0CAIBI0THrJrpnAAAg2ghmYSCcAQCAaGKUGSZGmwAAIFromEWI7hkAAOhpngWz6upqnXvuuSouLvaqhAMWTjgjoAEAgK54Mspcu3atbrvtNm3cuNGLr+9R3R5tTspRsGA5o00AANApTzpmixcv1sKFC5Wdne3F10cFo00AAHCgPOmY3XnnnV58bdTtu51TrslUfiBducFqbSjI0ojxFaFwxo0BAACgI9yV2cO6HG3uDWiMNgEAQFsEsyjpTvesJZxFilAHAEB8YbmMKOrourP8QLo2FGRpQ0GW6vLWhsabkeB6NQAA4gsdsyhrO9rc37VnkeB6NQAA4odjjDFeFxGOcTlTVLxpi9dlRGTsPuFMknKD1Qf0eSPGV0hqXopDIpwBAGALJylVKaMnhv0+Rpku2ne0me9UhkabkWIpDgAA4gujTJe1hLO2NwZEjD06Y4aTki4nI1tKSJSpq5bZtVUyQa/LAgBYhI6ZR9reGBApFrKNDU7v/lLGYJX+4y8qWXyvqr/4UAkHjZEc/hYEAHyNa8ws0HLtWSTaXq/GdWd2SjgkR8V/vFmmYU/oWJ9TzlfG4WMU3FXiYWUAgGiI9BozRpkWCN25GaaWcaikzvfo3PtaApqHElPUUL69VSiTpJp1bynjqLEeFQUAsBFzlBhWWFXEHp2xoKlBielZ7Q4n9R0kNTV6UBAAwFYEszhAOLOcCUpN9co88euuZSA1Xf3OvFymapuHhQEAbMMoM050uUcno01PBXcUKfOY05V5wlQF91QroXcfBXd+JVNf63VpAACL0DGLM3TPLGWMgjuKFNz6qbRrs5qKP5TZXeF1VQAAyxDM4hDhzGLGSEGuKwMAdIxRZpxitAkAQOyhYxbn6J4BABA7CGY+EE44I6ABAOAdRpk+0e3R5qQcBQuWM9oEAMADdMx8htEmAAD2Ipj5UHfCWV3eWkabAAC4jFGmT3U52hxfobq8tYw2AQBwER0zn2O0CQCAPeiYQYVVRRqbMVz5TqVyTWYonLHmGQAA7qJjBknN4aywqkj5TiXdMwAAPEIwQyuMNgEA8A6jTLSz72hTEts5AQDgEjpm6FDLaFOiewYAgFsIZtgvwhkAAO5hlIkudXs7p72vZ7QJAEBk6Jih2+ieAQAQXQQzhKVtOCur3SaJcAYAQE9glImw7TvanFyzXit7j2h+3MloMxKMQwEAfkQwQ8T27Z617BiQG6zWhoKs0F6bkeJ6NQCAHzHKxAGpaqqV1Pl1Z5FiJAoA8CM6ZjhgXd21GYmWjlvKpBwFC5bTOQPQJSfrIAXS+kmSgjVlMrtKPK4ICB/BDD2m083QI8FSHADCkDDoCFWvf18Vby2VJGWNm6Xeo45XcNvnHlcGhIdRJnpU27s2I8VSHAC6y0lJV0NVhcpXPSdTXytTX6vyVc+psapCTkpvr8sDwkLHDBFJTk7SmZNPV/brS3XIgKHa9NXm0HPtRpsR6OxmgpbRpkT3DEAzJy1TNR+/1e54zfp31WfMOJm6Gg+qAiJDMEPYjjhqhH7zxwf17vpdSl7zpu7/0yMqyMvT/b98qNXrQgEtEl3tMsC1ZwD2MnW71WvoEap6f2Wr470OPkKmfrdHVQGRYZSJsN3z27v1wIvr9eKq/6iiqk53PfOxck47QyecdGyPfQe7DADoLrO7Qr2GjlLaqJNCx9JGnaReB42S2V3hYWVA+OiYISwHDR2sqvqASstrWx1f/u42zbhwpt5798Me+66ObiboaJ00bgwA0FSyTv1Ou0D9z/y2JMnU71ZTyTqPqwLC5xhjjNdFhGNczhQVb9ridRlx7aK6cl1c3/F/ZSYnJ2vEkUdoU2nzNRvHFH+ijw4eo96pSeqlPfrPl191+rmLk7P015S+EdU0du9oM9c0jzZzg9WSmpfVkJpHmxLhDABgBycpVSmjJ4b/PoIZwvW/f3tGT6/aoeLt1bpr8a267ZI7ddOlY/SLH/1En378WdS+l3AGAIgVkQYzRpkI24+vuVmPPPVbbdhap/7Le+lnc3O05Nm/RjWUSftZJy0G1zxLWLpSTbMme10GAMAydMwQEcdxNHbcCbrr3TzNTj9MFeW7XP3+WO+eJc9foPrH7/a6DABAlETaMeOuTETEGKO3C9Zo585y10OZxF2bAID4xCgTMWvf0aakztc82/t6W7tnAAC0oGOGmFZYVUT3DAAQNwhmiAuEMwBAPGCUibjRbo9ORpsAgBhDxwxxx+bumZOaqUC/w6SkFCkxpes3AAB8hWCGuGRjOAv0H6YmJ0Vlb76oxqqdCgwYLqd3f9e+HwBgP0aZOCCLk7O8LqFTNo02nV4ZatxTq62Lm9cua6qp1NanbtXB8+9T0+6dUmwtJwgAiBI6Zjggke596SYbumdOWl9Vrmn9+aahTrX/+VhOap+ofS8AILaw8j98I9q7BSQsXamEl/M6fjIxWY01u9RU3fxdqRt3qHbYACUPPESOjBRs6vRzm2ZOYvsmAIgx7JUJdCHao82mWZM7D1CJyQoMHKWtT90qU1+rIX9cpR23fFuDL/iRmjZ/FNb3AADiF6NM+I4no83GepnKEh08/x4NnHGtkgce0hzKtq/vue8AAMQ8ghl8qTvhrC5vrYIFy3ssoJmanWra9KGSeyXLUbC5U9awp0c+GwAQHxhlwre6HG2Or1Bd3lqlTMpRsGB5D921aWRqK6VgsAc+CwAQb+iYwfdsuGsTnXNS0hXoO1SBjAFelwIAUUfHDFBzOBubMVz5TqVyTabyA+nKDVaHwhnbOXkjkD1SjTVVqv7gDaUMGqa0USeqqeTfUmOd16UBQFQQzIC9OhtttgS0tqPNSBHqusfJHKTa4vUqW/EnSVL1R2+o6pM3NWj6NWraus7j6gAgOghmQBv76561hLNIHWio8xMnNUsVb/2+1bH6kg1qXnjRkRRTSzACQLcQzIAO7BvOJHV4Y0AkWjpupnJnD1QZ74wcp4PLYB3H/VIAwCUEM6ATLaPN/V17FokRWquk8upQ54zRZsfM7nL1Pe1Clf7tkdCxXoeOkUyT6JYBiFdsyQR0Q9vtnCLVsg3U0UP+rfphgw94G6h4F+g/TEHjqOazQiVnH6peB41U09Z/S00NXpcGAPsV6ZZMLJcBdENhVZEKq4qU71R+Pd6MQMtSHOtKjpLEUhxdCZZtlKpKlDHyGCWnJjcvyksoAxDHGGUCYWh37VkkeniPzrjXsEfBXVu9rgIAXMEoE3BR25Foy2iz5WYCRpsAEB8YZQIxgF0GAAD7wygTcFmXe3Qy2gQA36JjBniE7hkAoC2CGeAhwhkAYF+MMgGPMdoEALSgYwZYgu4ZAKDLYNbY2Nju2K5du6JSDOB34YQzAhoAxJ9Og9nHH3+siRMn6vjjj9f111+v6urq0HNz5851ozbAl9ruMpAfSFd+IF0bCrK0oSBLdXlr6Z4BQJzqNJjdeeed+tnPfqZ//vOfSkxM1JVXXqn6+npJUoytSQvEJEabAOA/nQazPXv2aMKECerfv79+9atfKTs7W7fccoubtQG+151wVpe3ltEmAMSJToNZMBhUWVlZ6PE999yjL774Qr/73e/kOI4rxQHoerQp0T0DgHjRaTCbN2+ezjvvPL3xxhuSpNTUVD366KN68cUX9fnnn7tWIIBmjDYBIP51GsxmzZqlp556Sp988kno2EEHHaSXX36Zjhngkf2Fs31vDGC0CQCxab/LZQwfPlyvvvqqFi5cqPr6ehUXF2vu3Lk65ZRT3KoPQBudjTYlumcAEOu6XMfs+eefV2Njo2bPnq3LL79cF198sX7/+9+7URuA/WC0CQDxp8stmRzHUXJysmpraxUMBhljAhYprCrS2IzhoXDGdk4AENu67JjNnDlT1dXVWrp0qZ5++mk999xz+t73vudGbQC6oWW0KdE9A4BY12Uwu/baa3Xfffepd+/eOuyww7Ro0SINHz7cjdoAhIFwBgCxzzExtoz/uJwpKt60xesyAKuNzWj+j6dc0zzazA02b6k2YnyFJCllUo4kRpsAEC1OUqpSRk8M+31ddswAxB66ZwAQmwhmQJwinAFA7OnyrkwAsaslnGnvaJO7NgHAblxjBvhEd687ixShDgC+xjVmAParu6PNSDESBYADxygT8JHujjYjkTIpR8GC5XTOAOAAEMwAH9p3x4Bck6n8QLpyg9Wh7lkkuF4NAA4co0zApzobbUaKuz0B4MB50jF75ZVX9Oijj6qxsVFz5szR5Zdf7kUZgO91OtqMQEvHbcT4CtXlrQ2NNiW6ZwDQXa4Hs23btumBBx7Qiy++qOTkZF166aX65je/qZEjR7pdCoC92m2GHonOluLg2jMA6DbXR5mrV6/WuHHjlJWVpbS0NE2dOlWvvfaa22UAaCPUPYsQC9kCwIFzvWO2fft2DRw4MPQ4OztbH374odtlAOjAgYSzzm4maBltStwYAABdcT2YBYNBOY4TemyMafUYQGzqcimONteeEc4AoD3XR5mDBw9WaWlp6HFpaamys7PdLgNAlLBHJwBEzvVgdsopp6igoEA7d+5UbW2tVqxYodNPP93tMgBE0f7C2YaCLNXlrVVd3loFC5YT0ABgH66PMgcNGqQbbrhBV1xxhRoaGnThhRfq2GOPdbsMAFHW2Wizs2U1GG0CAJuYA3BBdzdQJ5wBiBeRbmLOlkwAoq7dOmmdrXm29/UENAB+xZZMAFxRWFXEjQEA0AWCGQBXEc4AoHOMMgG4rss1zxhtAvApOmYAPEP3DABaI5gB8BThDAC+xigTgOcYbQJAMzpmAKxB9wyA3xHMAFiFcAbAzxhlArAOo00AfkXHDIC16J4B8BuCGQCrdSec1eWtVbBgOQENQMxjlAnAel2ONsdXqC5vrVIm5ShYsJzRJoCYRccMQMxgtAkg3tExAxBTCquKNDZjuPKdSuWaTOUH0pUbrA6FM24MABDL6JgBiDmFVUUqrCpSvlPZ/CeQTvcMQFxwjDHG6yLCMS5nioo3bfG6DFjIcRxNmjpB55w3Xbt312rxn/9XH37widdlIcrG7r3uLNc0X3eWG6yW1HzdmSSlTMo5oM+n4wYgEk5SqlJGTwz/fQQzxIv7H71Xgb6H6v8+2K7UlESdf+pQvbJokZ554jmvS0OUtQ1nUs8GNMIZgHBFGsy4xgxx4bgTvqGMIYfrkaWfh47dV1yh2+d/W0sWL9XumloPq0O0tb1rs+21Zy13bUaK69UAuIVghrhw2hmn6Z3Pd7U6FgwaffzlLh1z7Gi9XbDGo8rgpq5uDIgES3EAcBMX/yMulG4vVf/MpHbHB/ZJVmlpmQcVwSv7uzEgEtxMAMBNdMwQF159aZleuOa7evezHdpRsUeSNHpYXyUHd6voi43eFgdPdNQ9ixh7dAJwCRf/I24cPfoI3X7/L9TgpKhXUkClJZt18/dvUUX5rq7fjLjVcmNApLq625NwBqAj3JUJ7NW3X5YaGhpUXVXjdSmwyIEENMIZgHBxVyawV/nOCq9LgIVCd25GorM9OhltAuhhXPwPAF1gj04AbiGYAUA3EM4AuIFRJgB0U9uFbBltAuhpdMwAIEx0zwBEC8HMA47jqE+fTCUkJHhdCoAIEc4ARAOjTJfNmn2urr7uKu2qbVRWeopefeEV/e7+//G6LAARYLQJoKfRMXPR+FNP1qVX/5d+9pdPdO9z63Tr4x9oxMkTNPeq73hdGoADQPcMQE8hmLlo/vev1JOvbVBDY1CSZIy0KG+jZl822+PKAByo7oSzury1ChYsJ6AB6BSjTBf1H9BPpRU7Wh1rbApKDvkYiAddjjbHV6gub61SJuUoWLCc0SaAdkgELnr3rTU66aj6BOW2AAAZBElEQVTsVscG909T+Y4dnbwDQCxitAkgUnTMXPTog4/pL0ueVFpKgj4qKtewwRmanTtUN151g9elAehhhVVFGpsxXPlOpXJNpvID6coNVofCGTcGAOgIHTMX7Swr12UzvqO6ogJdfHKqBtat1/xL5mvdp597XRqAKCisKlJhVZHyncrmP4F0umcA9ssxxhiviwjHuJwpKt60xesyACAsY/ded5Zrmq87yw1WS2q+7kySUiblSKJzBsQLJylVKaMnhv0+RpkA4IKORpuSWPMMQCuMMgHAJW1HmxI3BgBojWAGAC7jrk0AnWGUCQAeYDsnAB2hYwYAHqJ7BmBfBDMA8BjhDEALRpkAYAFGmwAkOmYAYJV9u2ejazbQPQN8ho4ZAFimZc2zoTXr9WnvEa22c2q7EXqk6LgBdiKYAYCFCquKVNVUG7rurNPR5t4dA8IVLFhOOAMsRDADAIvtbzP0lu5ZpLheDbAPwQwALLe/cBaptiNRwhlgBy7+B4AY0HY7p/xA+tf7bUaAmwkAO9ExA4AY0ulm6JFgKQ7AOnTMAMAyvXql6PDhhyoQ6Pgf0fsuqRG6OSACLMUB2IeOGQBY5HvXX6WZF87Upm3VGnPzlbrpsqt03+33t3tdSzhr6Z5FYn83E9A9A7xBMAMAS8y44BzlnD5Zt/1prYyRDt1areyjTtbcq76tJ3//dIfvCe0YEInOdhngxgDAM4wyAcASl8+7XM/+40sZ8/Wx5/+5Ueddcn5Uvo89OgH7OMbs+48A+43LmaLiTVu8LgMAInJRXbkurq/o8Llv5HxDG7fVhB4fU/yJPjp4jIYN7q2PP/h4v5+7ODlLf03pG1FNY/d2znJNc+csN1gdem7E+OZaWxaypXsGdI+TlKqU0RPDfx/BDADssPDu27TZOVRr/l0qSbpr8a166sePaNzQWl1/1Y1R//7OAhrhDAhfpMGMUSYAWOK39/1Os8YO1LRvHqxhQzKVlZGiOWcdrHt/fp8r389oE/AeF/8DgCV2lpXr4umX6YKLZ+rk445RY3W5Lj7nMlVVVXf95h7S6TpprHkGuIJRJgBY6q9VX+qijMM9+35Gm0DkGGUCAHoUo03AfYwyAQCdCq2T1tmaZ4w2gR5FxwwA0CW6Z4A7CGYAgG4hnAHRxygTANBtjDaB6KJjBgAIG90zIDoIZgBgqcXJWRG/t2+/LGVmZvRgNe0RzoCexygTACwVyd6XI0Ydrrt/c6eCialKTAiosmy7bv7+ApVuL4tChYw2gZ7GArMAECdSUpL1Ut4L+s2SL7Rt525J0qhD+ujiUwbowrMvjfr3syAt8DUWmAUAn5s87UwVfrYrFMokaf2mXdpR4+iYnNFR//5wRpuMN4GOMcoEgDgxMHuAdlQ3tDteuqteAwf2d6WGbo82J+UoWLCc7hnQBh0zAIgTq/Pf0thRfVodcxzpuJF99d67a12thRsDgMjQMQOAOPHZui+08dO1+q9zx2jFu1uVnJSgWaccpJf/9wVVVFS6Xk9hVZHGZgxXvlOpXJOp/EC6coPVoXDGjQFAe1z8DwBx5oxJp2nGhTNVV1ev559erPfe/dDrkrgxAL4T6cX/BDMAgCsIZ/CTSIMZo0wAgCs6Gm1KYs0zYB90zAAArutu9yxShDp4jVEmACCmdBXOIsVIFDZggVkfS+udqnNmnKVzZpyltN6pXpcDAN2y75Ia+U6l8gPpyg+ka0NBVujOzUiwFAdiGR2zGDfxrFz95Be36K11OyVJ447up3v+/1/q9X/ke1wZAHRfZ92zSHFDAbzGKNOH0nqn6sWVL+jnT32oPfVNkqReyQlaOOdYXTB5tnbX1HpcIQB0X9twFinu9oQNGGX60Blnnqq31u0MhTJJ2lPfpLfW7dQZZ57qYWUAEL7CqiIVVhWFRpuRYpcBxDKWy4h5HTU8jRzHcb0SAOgJ+y6rEbHO9uhseZruGSzFKDOG9U5P0wsrnu9wlDl7yoWqqd7tcYUA4D4WsoUNGGX6UE31bt2z8G4tnHOsZp56qGaeeqgWXnGM7ll4N6EMgG+xgTpiGaPMGPf6ylUqLHhXZ0w6TZL0q+vfJJQB8L2WcKa93TNGm4gVjDIBAHGN0Sa8wCgTAIAOMNpELGGUCQCIe4w2ESvomAEAfIPuGWxHMAMA+Eo44YyABrd5Nsp88MEHlZCQoB/84AdelQAA8KlujzYn5ShYsJzRJlzjesesqqpK//3f/60nnnjC7a8GAKAVRpuwjesds7y8PA0bNkzf/e533f5qAADa2XcLqFyTqfxAunKD1aFwxo0BcJPrHbPzzjtPV111lRISEtz+agAAOtR2A/X8QDrdM3giasFs2bJlOv3001v9mTt3brS+DgCAA8ZoE16L2ihz2rRpmjZtWrQ+HgCAqOhotCmJNc/gCpbLAACgjbajTYnuGdzh+2A2MLu/FvzsJi165Wn98sE7NHzkMK9LAgBYgtEm3ObrTcwHD8nWE399XEtWb9WHX5Tp8KGZuvzMw3TbdQv0wXsf9ch3AABiX8tG6BKboaN72MQ8Atfc8D0998ZmvbNuu+oamvTvjeV66MXPdNPCm7wuDfAlx3G8LgHoUMtoU6J7hujydTDLOeFYffhFWatjpRW1yszK8qgiwJ+OHnOkFr3yF73yxitavvpv+snCHysx0bONSYBOEc4Qbb7+J9+2rds1ZEBvleyoCR1LSUqQCTZ6WBXgLwMG9tf9j92vB1/8TNvLv5LjSNO+eYwW3n2bfvrjn3ldHtBOt7dz2vt6RpsIh687Zo/95jHNO3u40no159PEhIDmnD1CTz/+rMeVAf7xrTkX69XCEm0vr5UkGSP9/a1i5Zx8onqnp3lcHdA5umeIBl8Hs/feWatH7vm1br5wpH4+91j9fM4Y/evVF7X46ee9Lg3wjYMPO1TF23e3O769vFb9+vX1oCKg+whn6Gm+HmVK0v+teEP/t+INOY6jGLtBFYgL76x+W8dPmq2vtlWFjiUmOBo6IE1bNm/1sDKgexhtoif5erkMAN5LTk7Ss0v/rLeL6vSvj7ZpQFaqvnXmYXr+T0/ohede8ro8ICwty2p0taRGpAh1sSPS5TIIZgA8l5qWqu/M+5bOmDJRO3fs1JOPPal3337f67KAiHQVziLFOmmxhWAGAIBFohHQCGexI9Jg5vtrzAAAiIaONkPPDVaHbgyIBNerxT86ZgAARFHbzlmk2AIqtrAlEwAAFmrZzinfqQwtqxGJ/S3FwXIc8YNRJgAALth3tBmxzpbimJSjYMFyumdxgFEmAAAxoLtLcRDO7MDF/wAAxLGubibgxoD4wDVmAADEiLbXq+UH0tkGKs4QzAAAiDHs0Rm/GGUCABCDOhptSmKPzhhHxwwAgBjV0VIcdM9iG8EMAIAYx2gzfjDKBAAgDrRbJ62zNc/2vp7Rpp3omAEAECdaRpsS3bNYRTADACDOEM5iF6NMAADiUEs4094dAxhtxgY6ZgAAxDG6Z7GFYAYAQJwjnMUORpkAAPgAo83YQMcMAAAfoXtmN4IZAAA+E044I6C5i1EmAAA+1O3R5qQcBQuWM9p0CR0zAAB8jNGmXQhmAAD4XHfCWV3eWkabLmCUCQAAuh5tjq9QXd5aRptRRscMAACEMNr0Fh0zAADQSmFVkcZmDFe+U6lckxkKZ6x5Fn10zAAAQDuFVUUqrCpSvlNJ98xFjjHGeF1EOMblTFHxpi1elwEAgG+M3XvdWa5pvu4sN1gtqfm6M6l5SY0DEY8dNycpVSmjJ4b9PkaZAABgv/YdbUra75pnkeBmgq8RzAAAQJdabgpoe+1ZbrBaGwqyQndtRorr1ZoxygQAAGHpbLQZqbYj0XgIZ5GOMrn4HwAAhKXtjQH5gfSv79yMADcTfI1RJgAAiEiny2pEgqU4JDHKBAAAB6hltBmpru72jMVwxigTAAB4omW0GSnWSfsao0wAANAjDiScdbpHp89Gm3TMAACA59ijsxnBDAAAWIFwxigTAABYJDQO9elok44ZAACwjl+7ZwQzAABgpXDCWbwENEaZAADAWt0ebU7KiYvN0OmYAQAA6/lltEkwAwAAMaE74awub21MjzYZZQIAgJjR5WhzfIXq8tbG7GiTjhkAAIg58TrapGMGAABiUmFVkcZmDFe+U6lckxkKZ7G85hkdMwAAELNaNlDPdyrjontGMAMAADEvXkabjDIBAEBc2He0KSkmt3OiYwYAAOJGy2hTis3uGcEMAADEnVgNZ4wyAQBAXOr2dk57X2/DaJOOGQAAiGux1D0jmAEAgLgXK+GMUSYAAPCFWBht0jEDAAC+YnP3jGAGAAB8x9ZwxigTAAD4ko2jTTpmAADA12zqnhHMAACA73UnnNXlrVWwYHlUAxqjTAAAAHVjtDm+QnV5a5UyKUfBguVRGW3SMQMAANiHl6NNOmYAAABtFFYVaWzGcOU7lco1mcoPpCs3WB0KZ9G6MYCOGQAAQAcKq4pUWFWkfKey+U8gPerdM4IZAADAfrg52mSUCQAA0IV9R5uSulzzLOH08yL6HjpmAAAA3dAy2pS60T17f1VE30EwAwAACEN3w1kkGGUCAACEqas1z0YNXKfUy8P/XDpmAAAAEeqse/af9zIj+jyCGQAAwAHoLJxFglEmAADAAWo72swMDNCxEXwOHTMAAIAe0hLQCp3qiN5PMAMAAOhBoe5ZBAhmAAAAPWxt9VcRvY9gBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlXA9ma9as0YUXXqhZs2Zpzpw52rx5s9slAAAAWMn1YHbTTTfpjjvu0NKlSzVjxgzdcccdbpcAAABgJVeDWX19va677jodddRRkqQjjzxSJSUlbpYAAABgLVeDWXJysmbNmiVJCgaDevjhh3XWWWe5WQIAAIC1EqP1wcuWLdMvf/nLVseGDx+uJ598UvX19VqwYIEaGxt19dVXR6sEAACAmBK1YDZt2jRNmzat3fGamhpdc801ysrK0qOPPqqkpKRolQAAABBTPLn4/7DDDtODDz6o5ORkt78eAHrM6G8cqRnnn62RRwz3uhQAcSJqHbOOfPrpp8rLy9PIkSN1/vnnS5Kys7P1hz/8wc0yAOCA9ErtpUef+q2akrNUtK1WF87vrcptm3T9VTeqqanJ6/IAxDBXg9no0aP12WefufmVANDjbljwQ60tSdAbH3wuSVomacaph+q7V39Hf3zkSU9rAxDbWPkfAMJ06hmnatXa1kv9LH+7WOecP92jigDEC4IZAITJcZx2x4Km4+MAEA6CGQCE6e1/va3xYwa1Ojb5pIO08tUVHlUEIF64eo0ZAMSDX91xv/646DGNPmykvtxepyOG9FKgtkzX/O5PXpcGIMYRzAAgTLtranXZzCt04tjjNGLkMC3/5HN9tPZTr8sCEAcIZgAQoTWFH2hN4QdelwEgjnCNGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAAACWIJgBAABYgmAGAABgCYIZAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAb4VCAQUK9eKV6XAQDYR6LXBQBwV1JSohb87Gadesapqq1rUrChVnfeepfee3et16UBgO8RzACfWXj3T1Wddphu+eMHkqSsjBTd8eBdmn/JfJVs3upxdQDgb4wyAR9JTUtVzkknaHnh5tCxiqo6vbR6iy6be6mHlQEAJIIZ4CtZWZkqq9zT7viWHbs19JCDPagIALAvghngI9u2lio7K0XJia3/1j/piH4qWPUvj6oCALQgmAE+EgwG9dv7fqebv/UNjTw4S1npKZo27hCNOShBS59/1evyAMD3uPgf8JllLy/Xpo1fac5Vc9T/mAF6Y+UK3X/986qvb/C6NADwPYIZ4EMff7hON31/gddlAADaYJQJAABgCYIZAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAAACWIJgBAABYgmAGAABgCYIZAACAJRK9LiBcQw4a5HUJAAAA+xVpXnGMMaaHawEAAEAEGGUCAABYgmAGAABgCYIZAACAJQhmAAAAliCYAQAAWIJgBgAAYAmCGQAAgCUIZgAAAJYgmAEAAFiCYLbXli1bdPnll+vss8/WNddco5qamnav2bx5s44//njNmjVLs2bN0vz58z2oNLpeeeUVnXPOOZoyZYqeeeaZds+vW7dOF1xwgaZOnapbb71VjY2NHlTpnq5+j4cfflgTJ04MnRMdvSbeVFdX69xzz1VxcXG75/x2fkj7/z38dn48/PDDmj59uqZPn65777233fN+Oz+6+j38dn5I0m9+8xudc845mj59up544ol2z/vtHOmQgTHGmKuuusq8+uqrxhhjHn74YXPvvfe2e81rr71mfvrTn7pdmmu2bt1qJk6caMrLy01NTY2ZMWOGWb9+favXTJ8+3bz//vvGGGNuueUW88wzz3hRqiu683tcffXV5r333vOoQvd98MEH5txzzzVjxowxmzZtave8n84PY7r+Pfx0fvzrX/8yl1xyiamrqzP19fXmiiuuMCtWrGj1Gj+dH935Pfx0fhhjzNtvv20uvfRS09DQYGpra83EiRPNhg0bWr3GT+dIZ+iYSWpoaNA777yjqVOnSpIuuOACvfbaa+1e99FHH+nzzz/XrFmzdMUVV+izzz5zu9SoWr16tcaNG6esrCylpaVp6tSprX6HzZs3a8+ePTruuOMkdf47xYuufg9J+vjjj/XYY49pxowZ+sUvfqG6ujqPqnXH4sWLtXDhQmVnZ7d7zm/nh7T/30Py1/kxcOBALViwQMnJyUpKStKIESO0ZcuW0PN+Oz+6+j0kf50fkjR27Fj9+c9/VmJiosrKytTU1KS0tLTQ8347RzpDMJNUXl6u9PR0JSYmSmr+G2rbtm3tXpeSkqKZM2dqyZIlmj9/vq699lrV19e7XW7UbN++XQMHDgw9zs7ObvU7tH2+s98pXnT1e9TU1Ojoo4/WTTfdpCVLlqiyslKPPPKIF6W65s4779RJJ53U4XN+Oz+k/f8efjs/Ro0aFfoX6saNG7Vs2TJNmDAh9Lzfzo+ufg+/nR8tkpKS9NBDD2n69OkaP368Bg0aFHrOb+dIZ3wXzJYtW6bTTz+91Z8bb7xRjuO0el3bx5L0gx/8QJdddpkCgYAmTJigtLQ0FRUVuVV61AWDwVb/v40xrR539Xy86er/b+/evfWHP/xBI0aMUGJioubNm6c33njDi1Kt4Lfzoyt+PT/Wr1+vefPm6eabb9awYcNCx/16fnT2e/j1/JCkH/7whyooKFBJSYkWL14cOu7Xc6Qt3wWzadOmadWqVa3+/OlPf1JVVZWampokSaWlpR2OJv7yl7+ovLw89NgYE+qyxYPBgwertLQ09Ljt79D2+R07dnQ6wokHXf0eW7Zs0fPPPx96HG/nQ7j8dn50xY/nx5o1azR37lzdeOONOv/881s958fzY3+/hx/Pjw0bNmjdunWSpNTUVE2ZMqXVJUF+PEc64rtg1pGkpCSddNJJ+vvf/y5Jeumll3T66ae3e90777wT+hupsLBQwWBQw4cPd7XWaDrllFNUUFCgnTt3qra2VitWrGj1OwwdOlQpKSlas2aNJGnp0qUd/k7xoqvfo1evXrrvvvu0adMmGWP0zDPPaPLkyR5W7C2/nR9d8dv5UVJSomuvvVa/+tWvNH369HbP++386Or38Nv5IUnFxcW67bbbVF9fr/r6euXl5enEE08MPe+3c6RTntxyYKHi4mLz7W9/20ybNs3MmzfPVFRUGGOMefbZZ82DDz5ojGm+S2/u3Llm+vTp5oILLjDr1q3zsuSoePnll8306dPNlClTzO9//3tjjDFXXnml+fDDD40xxqxbt87Mnj3bTJ061fzoRz8ydXV1XpYbdV39Hq+99lro+QULFsT979Fi4sSJobsQ/Xx+tOjs9/DT+XH77beb4447zsycOTP059lnn/Xt+dGd38NP50eLhx56yEybNs2ce+655qGHHjLG8M+QthxjjPE6HAIAAIBRJgAAgDUIZgAAAJYgmAEAAFiCYAYAAGAJghkAAIAlCGYAoOYFPn/yk5/o8ccf97oUAD5GMAPgexs2bNCcOXO0fPlyr0sB4HMEMwC+sWTJEp111lmqqanR7t27NW3aNL300kt65plndNFFF+nss8/2ukQAPscCswB85cYbb1RGRobq6+uVkJCg22+/PfTcggULNGrUKM2fP9/DCgH4WXzvmAoAbfz85z/XrFmz1KtXL7344otelwMArTDKBOArZWVlqqurU2VlpbZv3+51OQDQCh0zAL7R0NCgH/3oR7ruuusUDAZ1ww03aNGiRUpKSvK6NACQRMcMgI/cf//9GjBggC666CJdcskl6tu3rx544AGvywKAEC7+BwAAsAQdMwAAAEsQzAAAACxBMAMAALAEwQwAAMASBDMAAABLEMwAAAAsQTADAACwBMEMAADAEv8PracSPcG4MMgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = svm.SVC(kernel='linear', C=1e5, random_state=0).fit(X, y)\n",
    "\n",
    "# Decistion boundary plot\n",
    "\n",
    "x1 = np.arange(-.7, 3.5, .1)\n",
    "x2 = np.arange(-2.5, 3, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df)\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500);\n",
    "\n",
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nWeights assigned to the features (coefficients in the primal problem).')\n",
    "print(model.coef_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "No training errors, only three support vectors used. We can see that margin is very narrorw because the support vectors are very close to the the decision boundary. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Now with a smaller value of cost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=1, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma='auto', kernel='linear',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[3 3]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[-0.90534066 -0.79575232 -1.          1.          0.70109298  1.        ]]\n",
      "\n",
      "Weights assigned to the features (coefficients in the primal problem).\n",
      "[[0.4497637  1.62175449]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = svm.SVC(kernel='linear', C=1, random_state=0).fit(X, y)\n",
    "\n",
    "# Decistion boundary plot\n",
    "\n",
    "x1 = np.arange(-.7, 3.5, .1)\n",
    "x2 = np.arange(-2.5, 3, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df)\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500);\n",
    "\n",
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nWeights assigned to the features (coefficients in the primal problem).')\n",
    "print(model.coef_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Using cost 1 we misclassify a training observation but there is a wider margin and so are more support vectors. It seems likely this model will be less affected by overfitting than when cost=1e5."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9.6.2 Support Vector Machine"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Geenrate data with a non-linear class boundary\n",
    "np.random.seed(1)\n",
    "X = np.random.normal(0, 1, (200, 2))\n",
    "X[1:100,]   = X[1:100,] + 2\n",
    "X[101:150,] = X[101:150,] - 2\n",
    "y = np.concatenate((np.repeat(0, 150), np.repeat(1, 50)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot data\n",
    "df = pd.concat([pd.DataFrame(data=X, columns=['x1', 'x2']), pd.Series(y, name='y')], axis=1)\n",
    "plt.figure(figsize=(10, 10))\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 255,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Index a training set\n",
    "train = np.random.random(len(y)) > 0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=1, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma=1, kernel='rbf',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[34 19]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[-0.37132271 -0.48342346 -0.33497995 -0.29851476 -0.0850208  -0.17263003\n",
      "  -1.         -0.17777894 -0.01244395 -0.6090602  -0.16824076 -0.31539116\n",
      "  -0.57799269 -0.08416399 -1.         -0.74614472 -1.         -0.25458311\n",
      "  -0.179926   -0.29396624 -0.210785   -0.56730368 -0.71857197 -0.97968356\n",
      "  -0.06553295 -0.2440551  -0.29817169 -0.79069468 -0.02047509 -1.\n",
      "  -0.693316   -0.12147924 -0.00789739 -0.77974867  1.          0.57788719\n",
      "   0.73784018  1.          1.          0.32380479  1.          0.38521508\n",
      "   1.          1.          1.          1.          1.          0.74624735\n",
      "   1.          0.57999088  0.30244074  0.00987226  1.        ]]\n",
      "\n",
      "Training accuracy:\n",
      "0.9553571428571429\n"
     ]
    }
   ],
   "source": [
    "model = svm.SVC(kernel='rbf', gamma=1, C=1, random_state=0).fit(X[train], y[train])\n",
    "\n",
    "# Decision boundary plot\n",
    "\n",
    "x1 = np.arange(-5, 5, .1)\n",
    "x2 = np.arange(-5, 5, .1)\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df[train])\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500)\n",
    "plt.show();\n",
    "\n",
    "# Get summary of model\n",
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nTraining accuracy:')\n",
    "print(model.score(X[train], y[train]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see from the figure that there are a fair number of training errors in this SVM fit. If we increase the value of cost, we can reduce the number of training errors. However, this comes at the price of a more irregular decision boundary that seems to be at risk of overfitting the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model parameters:\n",
      "<bound method BaseEstimator.get_params of SVC(C=100000.0, cache_size=200, class_weight=None, coef0=0.0,\n",
      "  decision_function_shape='ovr', degree=3, gamma=1, kernel='rbf',\n",
      "  max_iter=-1, probability=False, random_state=0, shrinking=True,\n",
      "  tol=0.001, verbose=False)>\n",
      "\n",
      "Number of support vectors for each class.:\n",
      "[13 12]\n",
      "\n",
      "Coefficients of the support vector in the decision function. :\n",
      "[[-2.94408546e+00 -1.28564417e+00 -3.46014897e+00 -1.10607541e+00\n",
      "  -4.38574425e+00 -1.07806443e+01 -2.95485465e+00 -1.25758441e+01\n",
      "  -2.56721287e+02 -4.33636718e-01 -7.63309008e+00 -3.08082873e+01\n",
      "  -1.50760502e+01  6.67359160e+00  1.30223908e+01  8.84018653e-01\n",
      "   2.84972061e+02  7.89188937e+00  8.98519034e+00  9.71729935e-01\n",
      "   2.50255629e+00  3.47333278e+00  1.68532851e+01  1.78243944e-01\n",
      "   3.75710246e+00]]\n",
      "\n",
      "Training accuracy:\n",
      "1.0\n"
     ]
    }
   ],
   "source": [
    "model = svm.SVC(kernel='rbf', gamma=1, C=1e5, random_state=0).fit(X[train], y[train])\n",
    "\n",
    "# Decision boundary plot\n",
    "\n",
    "x1 = np.arange(-5, 5, .1)\n",
    "x2 = np.arange(-5, 5, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "y_grid.shape\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contourf(x1, x2, y_grid);\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df[train])\n",
    "sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500)\n",
    "plt.show();\n",
    "\n",
    "# Get summary of model\n",
    "print('Model parameters:')\n",
    "print(model.get_params)\n",
    "\n",
    "print('\\nNumber of support vectors for each class.:')\n",
    "print(model.n_support_)\n",
    "\n",
    "print('\\nCoefficients of the support vector in the decision function. :')\n",
    "print(model.dual_coef_)\n",
    "\n",
    "print('\\nTraining accuracy:')\n",
    "print(model.score(X[train], y[train]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hyper-parameter tuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 310,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5, error_score='raise',\n",
       "       estimator=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),\n",
       "       fit_params=None, iid=True, n_jobs=1,\n",
       "       param_grid={'gamma': array([1.e-03, 1.e-02, 1.e-01, 1.e+00, 1.e+01, 1.e+02, 1.e+03, 1.e+04,\n",
       "       1.e+05, 1.e+06, 1.e+07, 1.e+08, 1.e+09]), 'C': array([1.e-02, 1.e-01, 1.e+00, 1.e+01, 1.e+02, 1.e+03, 1.e+04, 1.e+05,\n",
       "       1.e+06, 1.e+07, 1.e+08, 1.e+09, 1.e+10])},\n",
       "       pre_dispatch='2*n_jobs', refit=True, return_train_score='warn',\n",
       "       scoring=None, verbose=0)"
      ]
     },
     "execution_count": 310,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train Classifiers\n",
    "# https://scikit-learn.org/stable/auto_examples/svm/plot_rbf_parameters.html#sphx-glr-auto-examples-svm-plot-rbf-parameters-py\n",
    "# ----------------------------------------------------------------\n",
    "\n",
    "C_range     = np.logspace(-2, 10, 13)\n",
    "gamma_range = np.logspace(-3, 9, 13) \n",
    "param_grid  = dict(gamma=gamma_range, C=C_range)\n",
    "grid = GridSearchCV(svm.SVC(), param_grid=param_grid, cv=5)\n",
    "grid.fit(X[train], y[train])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 311,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The best parameters are {'C': 1.0, 'gamma': 1.0}, with a score of 0.929\n"
     ]
    }
   ],
   "source": [
    "print(f\"The best parameters are {grid.best_params_}, with a score of {grid.best_score_:.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Therefore, the best choice of parameters found involves cost=1 and gamma=1. We can view the test set predictions for this model by applying the predict() function to the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    Stolen from here: https://scikit-learn.org/stable/auto_examples/model_selection/plot_confusion_matrix.html#sphx-glr-auto-examples-model-selection-plot-confusion-matrix-py\n",
    "    \"\"\"\n",
    "    \n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        plt.title(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        plt.title('Confusion matrix, without normalization')\n",
    "        \n",
    "    #plt.title(title)\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "    plt.grid(False)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "    plt.tight_layout()\n",
    "    plt.show();\n",
    "\n",
    "cm = confusion_matrix(y[~train], grid.predict(X[~train]))\n",
    "plot_confusion_matrix(cm, ['True', 'False'], normalize=False)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10.23% of test observations are misclassified by this SVM.\n"
     ]
    }
   ],
   "source": [
    "print(f\"{(9 / len(y[~train]))*100:.2f}% of test observations are misclassified by this SVM.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9.6.3 ROC Curves"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### A single ROC curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 363,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Fit optimal model chosen by grid search \n",
    "model = svm.SVC(kernel='rbf', gamma=1, C=1, random_state=0, probability=True).fit(X[train], y[train])\n",
    "# Get probability of positive binary classification\n",
    "probs = model.predict_proba(X[~train])\n",
    "preds = probs[:, 1]\n",
    "# Get ROC metrics\n",
    "# False Postitive Rate, True Positive Rate metrics by threshold\n",
    "fpr, tpr, threshold = metrics.roc_curve(y[~train], preds)\n",
    "# Get area under curve metrics\n",
    "auc = metrics.auc(fpr, tpr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 364,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot ROC curve using seaborn\n",
    "plot_df = pd.DataFrame({'False Positive Rate': fpr, 'True Positive Rate': tpr})\n",
    "\n",
    "plt.figure(figsize=(10,10))\n",
    "sns.lineplot(x='False Positive Rate', y='True Positive Rate', data=plot_df, \n",
    "             estimator=None, \n",
    "             label=f'ROC curve (area={auc:.3f})');\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### ROC comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 501,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = np.empty((0, 3))\n",
    "for g in np.logspace(-5, 4, base=2, num=4):\n",
    "    # Fit optimal model chosen by grid search \n",
    "    model = svm.SVC(kernel='rbf', gamma=g, C=1, random_state=0, probability=True).fit(X[train], y[train])\n",
    "    # Get probability of positive binary classification\n",
    "    probs = model.predict_proba(X[~train])\n",
    "    preds = probs[:, 1]\n",
    "    # Get ROC metrics\n",
    "    # False Postitive Rate, True Positive Rate metrics by threshold\n",
    "    fpr, tpr, threshold = metrics.roc_curve(y[~train], preds)\n",
    "    # Get area under curve metrics\n",
    "    auc = metrics.auc(fpr, tpr)\n",
    "    r = np.array([np.repeat(g, len(fpr)), fpr, tpr]).T\n",
    "    results = np.concatenate((results, r), axis=0)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 502,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1a2295e4a8>"
      ]
     },
     "execution_count": 502,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAJVCAYAAACFyWIEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3X90VPWd//HXTCYJJoJgmDCUVgpmIRUTsXVtihYVlSAlqMguGpVtq1FK3RyptqJQ6GoRal3Qunps0Gp1oUrVgtl1AxWMFkNL4Vuk8qOIoCiQZEiAkJBIMnO/f6BT0yQEktyZdzLPxzk9h5nPnblv8ynm6Z1kxuM4jiMAAACY4o31AAAAAGiJSAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADCISAMAADDIF+sBOuPgwTqFw46r50hLO11VVbWungOnhj2xiX2xhz2xhz2xye198Xo96tcv9ZQf160jLRx2XI+0z84DW9gTm9gXe9gTe9gTmyzuCy93AgAAGESkAQAAGNStX+5sTSjUpIMHg2pqOtYlz1dZ6VU4HO6S57LG50tSv35+JST0uP8bAADQ7fW4784HDwbVq1eKUlMD8ng8nX4+n8+rpqaeF2mO46iurkYHDwbVv//AWI8DAAD+QY97ubOp6ZhSU/t0SaD1ZB6PR6mpfbrsiiMAAOhaPS7SJBFoJ4mvEwAAdvXISAMAAOjuiDQAAACDiDQAAACDetxvd1rz5JP/pdLS1TrjjL5KS+uviy8erY8+2qONG/+smpoa9e/fX/ffP19nnpmmiRNz9c1vXqKtW9/VmWf217e+NVEvvfSCgsFK3XffXJ1//td0xx23afjwTG3evEnHjh3TtGn/rt/+9gV98MEuTZmSrylTblQwWKn58x9Qbe0RHTgQ1Pjxebr11mmx/lIAAIBTwJU0F61d+5Y2b96k559fpp///FG9997fFAqFtGfPB3ryyV/phRde0YABAa1c+X+SpOrqKuXkjNIzzyzVsWOf6K233tATTzyl7373Ni1b9pvI8zqOo8WLn9Mll4zRI4/8XA8++HM98cRTeuaZpyRJv//9Sl15Za6Kip7Vc8+9qGXLfqNDhw7F5GsAAAA6hitpLtqw4U8aM+YKJSYmKjExUd/85iVKSEjQHXfMUHHxcu3Z86G2bPmrBg36YuQxOTkXSZICgYHKzh4pSRowIKAjR2paPWbEiCz16tVLgcBA1dYekSTl59+s//f/Nmjp0ue1e/f7ampqVENDvaS+UfonBwAAncWVNBd5vd4WH9h6+PBhzZhxhxwnrMsuu1yjR18qx/n7MYmJiZE/JyQktPq8Pp/vhMc89tgi/fa3LygQGKh/+7dbdMYZfZudAwAA2EekueiCC76uN99co8bGRtXV1aqsbK3q64/q/PO/pmuumawvfekslZWt7fKPndqw4U/Kz79ZY8ZcoT17PlQwWNljP9oKAICeipc7XTRq1MV6993N+s53blSfPn3Uv79fAwd+QevWva2pU6dIkoYP/4r279/Xpee96aZv64EH5ig5OVnp6QFlZp6jffv2NntZFQAA2OZxuvHrYFVVtS1eTiwv/1CBwOAuO0dnPrvz3Xc366OP9uiqqyaoqalJt9/+Hd177xxlZPxTl83XWV399YoGv7+3gsEjsR4D/4B9sYc9sYc9scntffF6PUpLO/2UH8eVNBedddZg/epXi/XCC0vkOGGNGzfBVKABAAC7iDQX9elzhhYufCzWYwAAgG7I9V8cqK2t1YQJE/Txxx+3WNu2bZsmTZqk3NxczZo1S01NTW6PAwAA0C24GmnvvPOObrjhBn3wwQetrv/whz/UnDlztHLlSjmOo2XLlrk5DgAAQLfh6sudy5Yt09y5c/WjH/2oxdrevXvV0NCgkSOPv2HrpEmT9Itf/EL5+flujgSgC3iOHpKnvqbZfTU1CfI2hmI0Uc9X/NparX7jz6f0GI/Hc3Lvkeg4x/+HqHBCYTm8LZIpE8ZfrCtuvCHWY7TgaqTNmzevzbXKykr5/f7Ibb/fr4qKilN6/tZ+U6Ky0iufr2svEHb181ni9Xrl9/eO9RinrDvO3JPUvL9XoaZjSjjttGb3Jya2/gbM6Lw1pRv0/vsfKyPjS6f0OI/H0+4xx4PBkdT+seg8JxyWwo7k5etticXvKzH7xYFwONzsXx6O45zUv0w+r7W34AiHwx1+y4zWdOYtOLqDcDjc7X4dnF9hjz1vY0jyJampz6DIfeyLu5yEJJ09/Gz9Z9GCk37Mye6JN/ihJCns715vx9Md+f29tfcPZZKkpMysGE+Dz1h9C46YXSIKBAIKBoOR2wcOHFB6enqsxnHNqlUluummf9H111+rl19u+TN37733N91yy826/vpJWrDgAX55AgAASIphpA0aNEjJycnauHGjJGnFihUaPXp0rMZxRTBYqcWLn9ATTzylZ55Zqldf/Z12797V7Jj77/+xZsz4kV544RU5jqPi4uUxmhYAAFgS9Zc7CwoKVFhYqKysLD388MOaPXu2amtrNWLECE2dOrVLz+U5elie+sOdeg7H45G3lR+odU47Q07KGSd87IYN6/XVr16gPn2OH3fZZZertHS1hgwZKkkqL9+vTz75ROeee/yS9/jxeXr66V/q2msnd2pmAADQ/UUl0tasWRP58+LFiyN/zszM1EsvvRSNEWLiwIGg0tL6R26npfXX1q1bTrheWVkZ1RkBAIBNPfoTB5yU9q92tcfn8yrUwV8caO2XI7yf+22e9tYBAED86rnvLWFAevoAVVUdiNyurq5S//7+k14HAADxi0hz0QUXXKiNG/+sgwcPqqGhQaWla/T1r38jsh4IDFRSUpI2b94kSSopeU05OaNiNS4AADCkR7/cGWt+f7oKCqarsPB2NTY2KS/vap1zzrm6++5C3XrrNGVmnqM5c36qhx76qerq6jRsWKYmT74+1mMDAAADiDSXjR07TmPHjmt238MP/yLy53/6p2FavPi5aI8FAACM4+VOAAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg3gLDpetXfuWnnlmsRoa6vXP/5yjO++8u9n6e+/9TQsWHH+ftJEjz9fdd98rn49tAQAg3nElzUV7936shx+er/nzH9avf/2CduzYrnXr3m52zP33/1gzZvxIL7zwihzHUXHx8hhNCwAALCHSXPTWW6W6/PIrlZ4+QD6fT/ffP18jRpwbWS8v369PPvlE556bJUkaPz5Pb7zxeqzGBQAAhvTo19UOb9+tQ9t2deo5PB6PHMdpcX/frwzVGZlDTvjYvXs/ks+XqHvumaGKigqNGnWxCgq+F1k/cCCotLT+kdtpaf1VWVnZqXkBAEDPwJU0F4VCIW3YsF4zZ87RL3/5K23d+q7+7//+J7IeDofl8Xgitx3Hkdfrae2pAABAnOnRV9LOyBzS7tWu9vh8XjU1hTv02DPPTNMFF1yofv36SZJGj75MW7du0fjxeZKk9PQBqqo6EDm+urpK/fv7OzUvAADoGXp0pMXaqFHf1Lx5c3XkyBGlpKToj38s0+jRl0TWA4GBSkpK0ubNm5SdPVIlJa8pJ2dUDCfuHE/dQXmO1rh+nppDCfI2hrr2ScNNUqipa5+zJ3PCkscrb/DDyF2f35fi19ZqdemG2IwWCskJdfH/PwzYtadcQ88KqHHzn0/6Mfu9Hjnhlj+u0cKn++lUuf/3N95Vvp8g52idPCmpsR4F3QAvd7poxIhzlZ8/VdOn36KbbvoXBQIBjR8/UXffXajt27dKkubM+akee2yh8vOvU339UU2efH2Mp+44z9EaqbEh1mN0TKjp+DcqnByPV0po+7/xVpdu0M5dH0dxoL9zQiHpZMKkmxl6VkCXfT3LnSf3eOWcYD/RtTwpqUo4k1dN0D6P09pPxXcTVVW1Cv/Dv4zLyz9UIDC4y87RmZc7u4Ou/Hp9dlUl7O+6r39r/P7eCgaPdOlzeqv2SJLCaWd16fPGk8/vyw9umylJWli0IOpzHNv+V0lSUqZLQdONuPF3BZ3Dntjk9r54vR6lpZ1+6o9zYRYAAAB0EpEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEWBXV1tbr55n/V/v37JEkrVryim2/+V02dOkUPPvgfamxsbPGY8vJyff/7BcrPv04zZ/5AR48ejfbYAAAghog0l23Z8q6mT79VH310/M1S9+z5UL/5zfN68slf6de/fkHhcFivvLKsxeMWLlyga6+drKVLX1Zm5jl69tmnoj06AACIoR79OSCb39ykTW/8pVPP4fFIrX0mw8jLzlf2JSPbfXxx8e/0gx/cowcemCNJSkpK0l133aPU1OPvPDx0aIYqKsqbPaapqUmbNv1FDz74sCTpqqsm6I47btf06YWd+mcBAADdR4+ONAtmzvxxs9uBwEAFAgMlSQcPHtQrryzTfffNbXbMoUOHlJqaKp/v+PakpfVXMFgRnYEBAIAJPTrSsi8ZeVJXu07Erc/uDAYrdffdhZow4Wp99asXNFtznLA8Hk+z+7xeXpkGACCe8J0/Bj788ANNm/ZdjRs3Qd/+9q0t1vv1O1O1tbUKhUKSpKqqA0pL80d7TAAAEENEWpQdPVqnGTO+r4KC7+mGG25q9Rifz6fzzhup1at/L0kqKflf5eSMiuaYAAAgxoi0KCsuXq6DB6v1wgtL9O1v5+vb387XU089KUlasOABrV37piTprrtm6tVXX9FNN/2L3nlnkwoKvhfLsQEAQJT16J9Js+Sll4olSVOm3KgpU25s9ZjP/5JBIDBQ//VfRVGZDQAA2MOVNAAAAIOINAAAAIOINAAAAIOINAAAAIP4xQE046k7KM/Rmo49uLFBSuzV4XOHKssVqg62e1zl+wlqbAyd4pM3yRM+wWOcsOTxygkePrXnjbHX1qxXadmmWI8hSfJ4PXLCxz9Dbdee/Rp61kAd2/7XqM/hHK2TJyU16ucFgK7GlTQ04zlaczy2OiKxl5yUPh0+d6g6KOdoXYcffyKecOh4iLV5gFeON8GVc7uptGyTdu3ZH+sxWhh61kBdOqpzn/bRUZ6UVCWcyZs/A+j+uJIWBXV1tZo27bt66KFHNHDgF7R+/R/1+OOPKhwOadiwTM2c+WMlJiY2e8yRI0d0//2ztW/fXvXt20/33z9faWn9ozNwYi+F/YOjc65/4ElJVVJm1gmP8ft7Kxg8ckrP663aI0kKp53V4dks8qakKiMzQwuLFsR6lA7tCwCgbVxJc9mWLe9q+vRb9dFHeyL3LVjwgP7jPx7U888v0yefNKik5H9bPG7x4ieUnX2+lix5SXl51+jRR/8zmmMDAIAYI9JcVlz8O/3gB/eof/+/v/wSCoVUV1enUCikY8eOKTk5ucXj1q17W2PHjpMkXXFFrv74xzI1NTVFbW4AABBbPfrlzlX/s1olr/6+U8/h8XjkOE6L+8dNvFJjJ1ze7uM//ykCn7nrrpn693+/XampqRo48Au69NKWz3PgQDDy8qbP51NqaqoOHTrYLPYAAEDPxZW0KKuqOqAnn3xMzz//olasKNGIEVl67LFFLY77xzB0HEcejydaYwIAgBjr0VfSxk64/KSudp2Iz+dVU9MJfivwFL3zziYNHXq2Bg36oiQpL+8azZlzb4vj/P50VVdXKT19gJqamnT06FGdcUbfLpsDAADYxpW0KBs69Gxt3bpF1dVVkqS1a9/UV75yTovjcnIuivxCwZo1v9d5542Uz9ejmxoAAHwO3/Wj7MtfHqKCgu+psHCaEhISNGjQl/SjH82SJD311JPq37+/rrlmsgoKpmnevJ/oppv+Vb17n645c34a48kBAEA0EWlR8tJLxZE/X3XVBF111YQWx9x667TIn/v0OUM/+1nLn1UDAADxgZc7AQAADCLSAAAADCLSAAAADOqRkdbam8+iJb5OAADY1eMizedLUl1dDQHSDsdxVFdXI58vKdajAACAVvS43+7s18+vgweDqq091CXP5/V6FQ533ZvZWuLzJalfPz5mCgAAi3pcpCUk+NS//8Auez6/v7eCwSNd9nwAAAAno8e93AkAANATEGkAAAAGEWkAAAAGEWkAAAAGEWkAAAAGEWkAAAAG9bi34OguPHUH5Tla486Th5ukUFPHHuuEJY9X3qo9XTuTpKaaWoVqj7Z96mPH5E1KavfcNTUJ8jaGTuncxcVvaPVbmyRf8ik9zrqdO3YrY9iQWI8BAHABV9JixHO0RmpscOfJQ03HY6sjPF4pwZ12D9UeVfjYsTbXvUlJSjg9xZVzr35rk3bu2uvKc8dSxrAhGjPu0liPAQBwAVfSYimxl8L+wV3+tJ9diQqnndXlz90ZTvCwPInJSszMOuFx7eVln468wbAvWRnDhmph0YJTexwAADHClTQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDiDQAAACDfLEeAF2vqaZWodqjcoKHo3re19asV2nZprYPCIckb4K8KamdOk9ikk+Nx5pO6TE7d+xWxrAhnTovAADRxJW0HihUe1ThY8eift7Ssk3atWd/2wd4E+RJTIzeQJ+TMWyIxoy7NCbnBgCgI7iS1kN5k5KUmJkV3XOmpCojM0MLixa4eh6/v7eCwSOungMAgFjjShoAAIBBRBoAAIBBRBoAAIBBrkZacXGxxo8fr7Fjx2rJkiUt1rds2aLrrrtOEydO1O23366amho3xwEAAOg2XIu0iooKLVq0SEuXLtXy5cv14osvaufOnc2OmTdvngoLC/Xqq69qyJAhevrpp90aBwAAoFtxLdLKysqUk5Ojvn37KiUlRbm5uSopKWl2TDgcVl1dnSSpvr5evXr1cmscAACAbsW1SKusrJTf74/cTk9PV0VFRbNjZs6cqdmzZ+viiy9WWVmZrr/+erfGAQAA6FZce5+0cDgsj8cTue04TrPbDQ0NmjVrlp599lllZ2frmWee0T333KOioqKTPkda2uldOnNb/P7eXf6cNYcSJEl9XHju/Z9+md2Y+0QSk3xRO2+0/9lwctgXe9gTe9gTmyzui2uRFggEtGHDhsjtYDCo9PT0yO0dO3YoOTlZ2dnZkqQpU6bo0UcfPaVzVFXVKhx2umbgNrj1xqnexpAkufLczqdfkmi/4etnH9Xk9nl5M1ub2Bd72BN72BOb3N4Xr9fToQtLrr3cOWrUKK1bt07V1dWqr6/XqlWrNHr06Mj64MGDVV5erl27dkmSVq9erays6L5DPgAAgFWuXUkbMGCAZsyYoalTp6qxsVGTJ09Wdna2CgoKVFhYqKysLM2fP1933nmnHMdRWlqaHnzwQbfGAQAA6FZc/ezOvLw85eXlNbtv8eLFkT9fcskluuSSS9wcAQAAoFviEwcAAAAMItIAAAAMItIAAAAMItIAAAAMItIAAAAMItIAAAAMItIAAAAMItIAAAAMcvXNbHEC4SYp1CRv1Z5TfmhTTa1CtUfbXP/f1/+k0g1b5U19qTMTnrKdO3YrY9iQqJ4TAICeiitpsRJqkpxwxx5ae1ThY8faXC/dsFW7Pqrs6GQdljFsiMaMuzTq5wUAoCfiSlosebwKp511yg9zgoflSUxWYmbrH0jvTX1JGZm9tbBoQWcnBAAAMcKVNAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIN8sR6gp/LUHZTnaE3bBzhhydN6I4cqyxWqDrb50NdK/qDSDdvkTUltdX3njt3KGDbklOYFAAC2cCXNJZ6jNVJjwwkO8EoJrTdyqDoo52hdmw8t3bBNuz4qb3M9Y9gQjRl36cmOCgAADOJKmpsSeynsH9zqkrdqzwkf6klJVVJmVuuPTUlVxvCztbBoQadHBAAANnElDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCAiDQAAwCBfrAforkIf71K4uqrtA5yw5PHKqappddnT+Mnxw4KHW6y9VvIHlW7YJm9KaquP3bljtzKGDTn1oQEAQLfBlbQOCldXKXysse0DPF45CR1r4NIN27Tro/I21zOGDdGYcZd26LkBAED3wJW0TvAmJSox+5879tiqPZKkcNpZLddSUpUx/GwtLFrQqfkAAED3xZU0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg4g0AAAAg3yxHiBeFb+2Vqvf+LPkS26xtnPHbmUMGxKDqQAAgBVcSYuR1W/8WTvf/7jVtYxhQzRm3KXRHQgAAJjClbQYyjj7i1pYtCDWYwAAAIO4kgYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGAQkQYAAGCQq5FWXFys8ePHa+zYsVqyZEmL9V27dunmm2/WxIkTdcstt+jw4cNujgMAANBtuBZpFRUVWrRokZYuXarly5frxRdf1M6dOyPrjuPoe9/7ngoKCvTqq6/qK1/5ioqKitwaBwAAoFtxLdLKysqUk5Ojvn37KiUlRbm5uSopKYmsb9myRSkpKRo9erQkadq0abrxxhvdGgcAAKBb8bn1xJWVlfL7/ZHb6enp2rx5c+T2nj171L9/f913333atm2bhg4dqh//+MendI60tNO7bN4T8ft7t7hvv9fT5trJ8Hg69/h4x9fNJvbFHvbEHvbEJov74lqkhcPhSIhIx1/e/PztpqYmrV+/Xv/93/+trKwsPfLII1qwYIEWLFhw0ueoqqpVOOx06dz/yO/vrWDwSIv7nU/P29rayXCczj0+nrW1J4gt9sUe9sQe9sQmt/fF6/V06MKSay93BgIBBYPByO1gMKj09PTIbb/fr8GDBysrK0uSNGHChGZX2gAAAOKZa5E2atQorVu3TtXV1aqvr9eqVasiP38mSeeff76qq6u1fft2SdKaNWs0YsQIt8YBAADoVlx7uXPAgAGaMWOGpk6dqsbGRk2ePFnZ2dkqKChQYWGhsrKy9Pjjj2v27Nmqr69XIBDQQw895NY4AAAA3YprkSZJeXl5ysvLa3bf4sWLI38+77zz9NJLL7k5AgAAQLfEJw4AAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAYRKQBAAAY1G6kNTY2RmMOAAAAfE67kTZ58uRozAEAAIDPaTfSTjvtNJWXl0djFgAAAHzK194B9fX1uvzyyxUIBJSSkhK5v7i42NXBAAAA4lm7kTZr1qxozAEAAIDPafflzgsvvFD9+vXT1q1btX37dqWnp+vCCy+MxmwAAABxq91I+81vfqOpU6dq69at2rx5s/Lz8/Xaa69FYzYAAIC41e7Lnc8++6yWL1+uAQMGSJL27dun2267TePHj3d9OAAAgHjV7pW0008/PRJokvSFL3xBSUlJrg4FAAAQ79q9knbRRRdp7ty5uvHGG5WQkKAVK1boy1/+srZs2SJJGjFihOtDAgAAxJt2I+2pp55SIBDQH/7wh8h99fX12rRpkzwej1avXu3qgAAAAPGozUg7dOiQJOnss8/W888/L8dx5PF41NjYqJtuukkrV66M2pAAAADxps1Iu+uuu/T222/L4/HoG9/4RuT+hIQE5ebmRmU4AACAeNVmpD399NOSpHvvvVfz58+P2kAAAAA4id/uJNAAAACir91IAwAAQPS1+9ud6BjP0UPy1Ne0fUA4LHlpZAAA0DoqwSWe+hqp8ZO2D/B6JW9C9AYCAADdClfS3JSYrHDaWa2v+ZKjOwsAAOhWuJIGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgkC/WA3RfjuQ48lbtaX258RMpMTm6IwEAgB6DK2kd5TgnXk9MlnNan+jMAgAAehyupHVSOO2sWI8AAAB6IK6kAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGESkAQAAGORqpBUXF2v8+PEaO3aslixZ0uZxpaWlGjNmjJujAAAAdCs+t564oqJCixYt0iuvvKKkpCRdf/31+vrXv66MjIxmxx04cEA/+9nP3BoDAACgW3LtSlpZWZlycnLUt29fpaSkKDc3VyUlJS2Omz17tu644w63xgAAAOiWXLuSVllZKb/fH7mdnp6uzZs3Nzvmuecnsi0eAAAQw0lEQVSe0znnnKPzzjuvQ+dISzu9UzOeLL+/d4v79nk8ba6djMQkX6ceH+/4utnEvtjDntjDnthkcV9ci7RwOCzPpyEjSY7jNLu9Y8cOrVq1Ss8++6zKy8s7dI6qqlqFw06nZz0Rv7+3gsEjLRec4+dtde0kNB5r6tTj41mbe4KYYl/sYU/sYU9scntfvF5Phy4sufZyZyAQUDAYjNwOBoNKT0+P3C4pKVEwGNR1112n2267TZWVlcrPz3drHAAAgG7FtUgbNWqU1q1bp+rqatXX12vVqlUaPXp0ZL2wsFArV67UihUrVFRUpPT0dC1dutStcQAAALoV1yJtwIABmjFjhqZOnaprrrlGEyZMUHZ2tgoKCvTXv/7VrdMCAAD0CK79TJok5eXlKS8vr9l9ixcvbnHcF7/4Ra1Zs8bNUQAAALoVPnEAAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAICINAADAIF+sB+ip/ueVEq0pKW1zfeeO3coYNiR6AwEAgG6FK2kuWVNSqp07dre5njFsiMaMuzR6AwEAgG6FK2kuyhg2RAuLFsR6DAAA0A1xJQ0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgIg0AAMAgX6wHsCxUWa7K97eqsTHUYs1pDMmbmBCDqQAAQDzgStoJhKqDaqw50uqaNzFBCSm9ojwRAACIF1xJa0din97ynH1Oi/u9+/4mSQpHeyAAABAXuJIGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgEJEGAABgkKuRVlxcrPHjx2vs2LFasmRJi/XXX39dV199tSZOnKjp06fr8OHDbo4DAADQbbgWaRUVFVq0aJGWLl2q5cuX68UXX9TOnTsj67W1tfrJT36ioqIivfrqqxo+fLgee+wxt8YBAADoVlyLtLKyMuXk5Khv375KSUlRbm6uSkpKIuuNjY2aO3euBgwYIEkaPny49u/f79Y4AAAA3YprkVZZWSm/3x+5nZ6eroqKisjtfv366corr5QkNTQ0qKioSFdccYVb4wAAAHQrPreeOBwOy+PxRG47jtPs9meOHDmi73//+8rMzNS11157SudISzu903OeSOX7CZIkv793i7Xqck+ba5KUmOQ74To6h6+rTeyLPeyJPeyJTRb3xbVICwQC2rBhQ+R2MBhUenp6s2MqKyt1yy23KCcnR/fdd98pn6OqqlbhsNPpWdvS2BhSYmKCgsEjLda8n563tTVJajzWdMJ1dJzf35uvq0Hsiz3siT3siU1u74vX6+nQhSXXXu4cNWqU1q1bp+rqatXX12vVqlUaPXp0ZD0UCmnatGm66qqrNGvWrFavsgEAAMQr166kDRgwQDNmzNDUqVPV2NioyZMnKzs7WwUFBSosLFR5ebm2bt2qUCiklStXSpLOPfdczZs3z62RAAAAug3XIk2S8vLylJeX1+y+xYsXS5KysrK0fft2N08PAADQbfGJAwAAAAYRaQAAAAYRaQAAAAa5+jNpPVlxyTqtfusvUtJpra7v3LFbGcOGRHkqAADQU3AlrYNWv/UX7dy9t831jGFDNGbcpdEbCAAA9ChcSeuEjCGDtLBoQazHAAAAPRBX0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAzyxXoAy15bs15v/vEdeXqltFjbuXuvMoYMisFUAAAgHnAl7QRKyzbp/Q/2tbqWMWSQLh99fpQnAgAA8YIrae04+8tf0MKiBS3u9+77myQpHO2BAABAXOBKGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEFEGgAAgEG+WA9gmyOFwvIGP2xlKSx5aFwAAOAOKuNEHEdOW2ser5RA4wIAAHdQGe3wSAr7B7e431u1J/rDAACAuMGVNAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIOINAAAAIN8sR7AOkeSt2pPy4XGT6TE5KjPAwAA4gORdgJXjR0ljxNqfTExWc5pfaI7EAAAiBtE2gnk/tuN8vt7Kxg8EutRAABAnOFn0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAwi0gAAAAxyNdKKi4s1fvx4jR07VkuWLGmxvm3bNk2aNEm5ubmaNWuWmpqa3BwHAACg23At0ioqKrRo0SItXbpUy5cv14svvqidO3c2O+aHP/yh5syZo5UrV8pxHC1btsytcQAAALoV1yKtrKxMOTk56tu3r1JSUpSbm6uSkpLI+t69e9XQ0KCRI0dKkiZNmtRsHQAAIJ753HriyspK+f3+yO309HRt3ry5zXW/36+KiopTOkda2umdH/Qk+P29o3IenDz2xCb2xR72xB72xCaL++LalbRwOCyPxxO57ThOs9vtrQMAAMQz1yItEAgoGAxGbgeDQaWnp7e5fuDAgWbrAAAA8cy1SBs1apTWrVun6upq1dfXa9WqVRo9enRkfdCgQUpOTtbGjRslSStWrGi2DgAAEM88juM4bj15cXGxfvnLX6qxsVGTJ09WQUGBCgoKVFhYqKysLG3fvl2zZ89WbW2tRowYofnz5yspKcmtcQAAALoNVyMNAAAAHcMnDgAAABhEpAEAABhEpAEAABhEpAEAABhEpAEAABhEpOn4W4WMHz9eY8eO1ZIlS1qsb9u2TZMmTVJubq5mzZqlpqamGEwZf9rbl9dff11XX321Jk6cqOnTp+vw4cMxmDK+tLcnnyktLdWYMWOiOFl8a29fdu3apZtvvlkTJ07ULbfcwt+VKGhvT7Zs2aLrrrtOEydO1O23366ampoYTBl/amtrNWHCBH388cct1kx+r3fiXHl5uXPZZZc5Bw8edOrq6py8vDznvffea3bMt771Lecvf/mL4ziOc++99zpLliyJxahxpb19OXLkiHPRRRc55eXljuM4ziOPPOI88MADsRo3LpzM3xXHcZxgMOiMGzfOueyyy2IwZfxpb1/C4bAzduxY580333Qcx3F+/vOfOw899FCsxo0LJ/N35YYbbnBKS0sdx3Gc+fPnOwsXLozFqHFl06ZNzoQJE5wRI0Y4H330UYt1i9/r4/5KWllZmXJyctS3b1+lpKQoNzdXJSUlkfW9e/eqoaFBI0eOlCRNmjSp2Trc0d6+NDY2au7cuRowYIAkafjw4dq/f3+sxo0L7e3JZ2bPnq077rgjBhPGp/b2ZcuWLUpJSYl8osu0adN04403xmrcuHAyf1fC4bDq6uokSfX19erVq1csRo0ry5Yt09y5c1v9CEqr3+vjPtIqKyvl9/sjt9PT01VRUdHmut/vb7YOd7S3L/369dOVV14pSWpoaFBRUZGuuOKKqM8ZT9rbE0l67rnndM455+i8886L9nhxq7192bNnj/r376/77rtP1157rebOnauUlJRYjBo3TubvysyZMzV79mxdfPHFKisr0/XXXx/tMePOvHnzdMEFF7S6ZvV7fdxHWjgclsfjidx2HKfZ7fbW4Y6T/bofOXJEt912mzIzM3XttddGc8S4096e7NixQ6tWrdL06dNjMV7cam9fmpqatH79et1www363e9+py996UtasGBBLEaNG+3tSUNDg2bNmqVnn31Wa9euVX5+vu65555YjIpPWf1eH/eRFggEFAwGI7eDwWCzS6H/uH7gwIFWL5Wia7W3L9Lx//LJz8/X8OHDNW/evGiPGHfa25OSkhIFg0Fdd911uu222yL7A3e1ty9+v1+DBw9WVlaWJGnChAnavHlz1OeMJ+3tyY4dO5ScnKzs7GxJ0pQpU7R+/fqoz4m/s/q9Pu4jbdSoUVq3bp2qq6tVX1+vVatWRX52Q5IGDRqk5ORkbdy4UZK0YsWKZutwR3v7EgqFNG3aNF111VWaNWuWif/i6ena25PCwkKtXLlSK1asUFFRkdLT07V06dIYThwf2tuX888/X9XV1dq+fbskac2aNRoxYkSsxo0L7e3J4MGDVV5erl27dkmSVq9eHYloxIbV7/W+WA8QawMGDNCMGTM0depUNTY2avLkycrOzlZBQYEKCwuVlZWlhx9+WLNnz1Ztba1GjBihqVOnxnrsHq+9fSkvL9fWrVsVCoW0cuVKSdK5557LFTUXnczfFUTfyezL448/rtmzZ6u+vl6BQEAPPfRQrMfu0U5mT+bPn68777xTjuMoLS1NDz74YKzHjkvWv9d7HMdxYj0EAAAAmov7lzsBAAAsItIAAAAMItIAAAAMItIAAAAMItIAAAAMItIAxKW5c+dqzJgxWrRoUaxHAYBW8RYcAOJSZmamSktLFQgEYj0KALSKSAMQd/Lz87Vx40YNGzZMO3fu1K233qq3335bR44c0Xe+8x3l5+frT3/6k+bNm6eUlBTV1dXp5ZdfVlJSUqxHBxBH4v4TBwDEn6VLl2r48OH69a9/rcmTJ+vw4cN6+eWXVVFRoWuuuUZf+9rXJEnvvfeeXn/9dQ0aNCjGEwOIR/xMGoC4l5+fL4/Ho0AgoG9+85t6++23JUkDBw4k0ADEDJEGIO75fH9/USEcDsvrPf6vxpSUlFiNBABEGgAsX75ckrRv3z69/fbbGj16dIwnAgB+Jg0A9PHHH2vSpElqaGjQ7NmzNXToUAWDwViPBSDO8dudAOLamDFj9OijjyorKyvWowBAM7zcCQAAYBBX0gAAAAziShoAAIBBRBoAAIBBRBoAAIBBRBoAAIBBRBoAAIBBRBoAAIBB/x9g7V/1BCM4TwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_df = pd.DataFrame(results, columns=['gamma', 'fpr', 'tpr'])\n",
    "plt.figure(figsize=(10,10))\n",
    "sns.lineplot(x='fpr', y='tpr', hue='gamma', data=plot_df, estimator=None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 9.6.4 SVM with Multiple Classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 539,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate multi-class data with 3 response classes 0, 1 and 2\n",
    "np.random.seed(0)\n",
    "X = np.random.normal(0, 1, (300, 2))\n",
    "X[1:100,]   = X[1:100,]\n",
    "X[101:200, :1] = X[101:200, :1] + 4\n",
    "X[201:300, 1:] = X[201:300, 1:] + 4\n",
    "X[201:300, :1] = X[201:300, :1] + 2\n",
    "y = np.concatenate((np.repeat(0, 100), np.repeat(1, 100), np.repeat(2, 100)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 540,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot data\n",
    "df = pd.concat([pd.DataFrame(data=X, columns=['x1', 'x2']), pd.Series(y, name='y')], axis=1)\n",
    "plt.figure(figsize=(10, 10))\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 570,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = svm.SVC(kernel='rbf', gamma=1, C=1, random_state=0, probability=True).fit(X, y)\n",
    "\n",
    "# Decision boundary plot\n",
    "\n",
    "x1 = np.arange(-5, 7, .1)\n",
    "x2 = np.arange(-5, 7, .1)\n",
    "\n",
    "\n",
    "xx1, xx2 = np.meshgrid(x1, x2, sparse=False)\n",
    "\n",
    "Xgrid = np.stack((xx1.flatten(), xx2.flatten())).T\n",
    "y_hat = model.predict(Xgrid)\n",
    "y_grid = y_hat.reshape(len(x2), len(x1))\n",
    "\n",
    "fig = plt.figure(figsize=(10, 10))\n",
    "plt.contour(x1, x2, y_grid);\n",
    "\n",
    "sns.scatterplot(x='x1', y='x2', hue='y', data=df)\n",
    "#sns.scatterplot(x=model.support_vectors_[:,0], y=model.support_vectors_[:,1], color='red', marker='+', s=500)\n",
    "plt.show();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9.6.5 Application to Gene Expression Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 584,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>EWS.T1</th>\n",
       "      <th>EWS.T2</th>\n",
       "      <th>EWS.T3</th>\n",
       "      <th>EWS.T4</th>\n",
       "      <th>EWS.T6</th>\n",
       "      <th>EWS.T7</th>\n",
       "      <th>EWS.T9</th>\n",
       "      <th>EWS.T11</th>\n",
       "      <th>EWS.T12</th>\n",
       "      <th>EWS.T13</th>\n",
       "      <th>...</th>\n",
       "      <th>RMS.T4</th>\n",
       "      <th>RMS.T2</th>\n",
       "      <th>RMS.T6</th>\n",
       "      <th>RMS.T7</th>\n",
       "      <th>RMS.T8</th>\n",
       "      <th>RMS.T5</th>\n",
       "      <th>RMS.T9</th>\n",
       "      <th>RMS.T3</th>\n",
       "      <th>RMS.T10</th>\n",
       "      <th>RMS.T11</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3.2025</td>\n",
       "      <td>1.6547</td>\n",
       "      <td>3.2779</td>\n",
       "      <td>1.0060</td>\n",
       "      <td>2.7098</td>\n",
       "      <td>2.0588</td>\n",
       "      <td>1.8483</td>\n",
       "      <td>2.7140</td>\n",
       "      <td>2.3555</td>\n",
       "      <td>1.9291</td>\n",
       "      <td>...</td>\n",
       "      <td>3.4636</td>\n",
       "      <td>2.0816</td>\n",
       "      <td>3.1013</td>\n",
       "      <td>2.0272</td>\n",
       "      <td>2.2313</td>\n",
       "      <td>1.8594</td>\n",
       "      <td>2.5447</td>\n",
       "      <td>1.2705</td>\n",
       "      <td>1.2766</td>\n",
       "      <td>2.0298</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0681</td>\n",
       "      <td>0.0710</td>\n",
       "      <td>0.1160</td>\n",
       "      <td>0.1906</td>\n",
       "      <td>0.2367</td>\n",
       "      <td>0.0823</td>\n",
       "      <td>0.1234</td>\n",
       "      <td>0.1805</td>\n",
       "      <td>0.0792</td>\n",
       "      <td>0.2520</td>\n",
       "      <td>...</td>\n",
       "      <td>1.2855</td>\n",
       "      <td>0.9137</td>\n",
       "      <td>0.3910</td>\n",
       "      <td>0.5502</td>\n",
       "      <td>1.9247</td>\n",
       "      <td>0.5240</td>\n",
       "      <td>0.5169</td>\n",
       "      <td>0.4657</td>\n",
       "      <td>0.7770</td>\n",
       "      <td>0.7067</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0460</td>\n",
       "      <td>1.0409</td>\n",
       "      <td>0.8926</td>\n",
       "      <td>0.4302</td>\n",
       "      <td>0.3693</td>\n",
       "      <td>0.9021</td>\n",
       "      <td>0.9983</td>\n",
       "      <td>0.4964</td>\n",
       "      <td>0.7614</td>\n",
       "      <td>0.5745</td>\n",
       "      <td>...</td>\n",
       "      <td>0.3355</td>\n",
       "      <td>0.5806</td>\n",
       "      <td>0.3937</td>\n",
       "      <td>0.3688</td>\n",
       "      <td>0.2943</td>\n",
       "      <td>0.6808</td>\n",
       "      <td>1.2190</td>\n",
       "      <td>0.9344</td>\n",
       "      <td>0.2212</td>\n",
       "      <td>1.0439</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.1243</td>\n",
       "      <td>0.0520</td>\n",
       "      <td>0.1014</td>\n",
       "      <td>0.1035</td>\n",
       "      <td>0.2190</td>\n",
       "      <td>0.1288</td>\n",
       "      <td>0.2203</td>\n",
       "      <td>0.2509</td>\n",
       "      <td>0.1868</td>\n",
       "      <td>0.1356</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0893</td>\n",
       "      <td>0.0673</td>\n",
       "      <td>0.2905</td>\n",
       "      <td>0.3627</td>\n",
       "      <td>0.1762</td>\n",
       "      <td>0.3013</td>\n",
       "      <td>0.5151</td>\n",
       "      <td>0.0680</td>\n",
       "      <td>0.1432</td>\n",
       "      <td>0.1016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.4941</td>\n",
       "      <td>0.2045</td>\n",
       "      <td>0.2818</td>\n",
       "      <td>0.2984</td>\n",
       "      <td>0.3711</td>\n",
       "      <td>0.3961</td>\n",
       "      <td>0.3766</td>\n",
       "      <td>0.4754</td>\n",
       "      <td>0.4167</td>\n",
       "      <td>0.3363</td>\n",
       "      <td>...</td>\n",
       "      <td>0.3219</td>\n",
       "      <td>0.6856</td>\n",
       "      <td>0.1113</td>\n",
       "      <td>0.6133</td>\n",
       "      <td>0.3855</td>\n",
       "      <td>0.4011</td>\n",
       "      <td>0.3405</td>\n",
       "      <td>0.3930</td>\n",
       "      <td>0.1669</td>\n",
       "      <td>0.2147</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 64 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   EWS.T1  EWS.T2  EWS.T3  EWS.T4  EWS.T6  EWS.T7  EWS.T9  EWS.T11  EWS.T12  \\\n",
       "1  3.2025  1.6547  3.2779  1.0060  2.7098  2.0588  1.8483   2.7140   2.3555   \n",
       "2  0.0681  0.0710  0.1160  0.1906  0.2367  0.0823  0.1234   0.1805   0.0792   \n",
       "3  1.0460  1.0409  0.8926  0.4302  0.3693  0.9021  0.9983   0.4964   0.7614   \n",
       "4  0.1243  0.0520  0.1014  0.1035  0.2190  0.1288  0.2203   0.2509   0.1868   \n",
       "5  0.4941  0.2045  0.2818  0.2984  0.3711  0.3961  0.3766   0.4754   0.4167   \n",
       "\n",
       "   EWS.T13   ...     RMS.T4  RMS.T2  RMS.T6  RMS.T7  RMS.T8  RMS.T5  RMS.T9  \\\n",
       "1   1.9291   ...     3.4636  2.0816  3.1013  2.0272  2.2313  1.8594  2.5447   \n",
       "2   0.2520   ...     1.2855  0.9137  0.3910  0.5502  1.9247  0.5240  0.5169   \n",
       "3   0.5745   ...     0.3355  0.5806  0.3937  0.3688  0.2943  0.6808  1.2190   \n",
       "4   0.1356   ...     0.0893  0.0673  0.2905  0.3627  0.1762  0.3013  0.5151   \n",
       "5   0.3363   ...     0.3219  0.6856  0.1113  0.6133  0.3855  0.4011  0.3405   \n",
       "\n",
       "   RMS.T3  RMS.T10  RMS.T11  \n",
       "1  1.2705   1.2766   2.0298  \n",
       "2  0.4657   0.7770   0.7067  \n",
       "3  0.9344   0.2212   1.0439  \n",
       "4  0.0680   0.1432   0.1016  \n",
       "5  0.3930   0.1669   0.2147  \n",
       "\n",
       "[5 rows x 64 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>TEST.9</th>\n",
       "      <th>TEST.11</th>\n",
       "      <th>TEST.5</th>\n",
       "      <th>TEST.8</th>\n",
       "      <th>TEST.10</th>\n",
       "      <th>TEST.13</th>\n",
       "      <th>TEST.3</th>\n",
       "      <th>TEST.1</th>\n",
       "      <th>TEST.2</th>\n",
       "      <th>TEST.4</th>\n",
       "      <th>...</th>\n",
       "      <th>TEST.20</th>\n",
       "      <th>TEST.17</th>\n",
       "      <th>TEST.18</th>\n",
       "      <th>TEST.22</th>\n",
       "      <th>TEST.16</th>\n",
       "      <th>TEST.23</th>\n",
       "      <th>TEST.14</th>\n",
       "      <th>TEST.25</th>\n",
       "      <th>TEST.15</th>\n",
       "      <th>TEST.19</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.7733</td>\n",
       "      <td>0.1397</td>\n",
       "      <td>1.9420</td>\n",
       "      <td>0.7721</td>\n",
       "      <td>0.3296</td>\n",
       "      <td>0.7509</td>\n",
       "      <td>1.4407</td>\n",
       "      <td>1.1497</td>\n",
       "      <td>3.2036</td>\n",
       "      <td>2.3189</td>\n",
       "      <td>...</td>\n",
       "      <td>3.0381</td>\n",
       "      <td>3.7487</td>\n",
       "      <td>0.1860</td>\n",
       "      <td>0.7623</td>\n",
       "      <td>0.8264</td>\n",
       "      <td>0.6403</td>\n",
       "      <td>0.6729</td>\n",
       "      <td>0.8249</td>\n",
       "      <td>0.1181</td>\n",
       "      <td>0.7173</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.4875</td>\n",
       "      <td>0.0846</td>\n",
       "      <td>0.2103</td>\n",
       "      <td>0.1855</td>\n",
       "      <td>0.3510</td>\n",
       "      <td>0.4165</td>\n",
       "      <td>0.3036</td>\n",
       "      <td>0.3107</td>\n",
       "      <td>0.1329</td>\n",
       "      <td>1.2901</td>\n",
       "      <td>...</td>\n",
       "      <td>0.1378</td>\n",
       "      <td>1.6411</td>\n",
       "      <td>0.0872</td>\n",
       "      <td>0.3245</td>\n",
       "      <td>0.2202</td>\n",
       "      <td>0.2515</td>\n",
       "      <td>0.3038</td>\n",
       "      <td>0.3454</td>\n",
       "      <td>0.1068</td>\n",
       "      <td>0.1108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.5832</td>\n",
       "      <td>2.3266</td>\n",
       "      <td>2.4897</td>\n",
       "      <td>1.1922</td>\n",
       "      <td>0.4258</td>\n",
       "      <td>1.2165</td>\n",
       "      <td>2.9190</td>\n",
       "      <td>1.7594</td>\n",
       "      <td>3.0148</td>\n",
       "      <td>0.8116</td>\n",
       "      <td>...</td>\n",
       "      <td>1.2505</td>\n",
       "      <td>1.0424</td>\n",
       "      <td>1.9834</td>\n",
       "      <td>0.2148</td>\n",
       "      <td>2.2806</td>\n",
       "      <td>1.3489</td>\n",
       "      <td>2.0071</td>\n",
       "      <td>1.2253</td>\n",
       "      <td>0.3881</td>\n",
       "      <td>0.6291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.3514</td>\n",
       "      <td>0.1188</td>\n",
       "      <td>0.1770</td>\n",
       "      <td>0.0979</td>\n",
       "      <td>0.0737</td>\n",
       "      <td>0.1280</td>\n",
       "      <td>1.1618</td>\n",
       "      <td>0.0345</td>\n",
       "      <td>0.1147</td>\n",
       "      <td>0.1167</td>\n",
       "      <td>...</td>\n",
       "      <td>0.2309</td>\n",
       "      <td>0.2140</td>\n",
       "      <td>0.1222</td>\n",
       "      <td>0.0707</td>\n",
       "      <td>0.0915</td>\n",
       "      <td>0.0646</td>\n",
       "      <td>0.1553</td>\n",
       "      <td>0.1277</td>\n",
       "      <td>0.0622</td>\n",
       "      <td>0.0742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.7182</td>\n",
       "      <td>0.3497</td>\n",
       "      <td>0.5963</td>\n",
       "      <td>0.1841</td>\n",
       "      <td>0.1702</td>\n",
       "      <td>0.2474</td>\n",
       "      <td>0.1502</td>\n",
       "      <td>0.2663</td>\n",
       "      <td>0.2369</td>\n",
       "      <td>0.2203</td>\n",
       "      <td>...</td>\n",
       "      <td>0.3002</td>\n",
       "      <td>0.1965</td>\n",
       "      <td>0.1836</td>\n",
       "      <td>0.3171</td>\n",
       "      <td>0.3287</td>\n",
       "      <td>0.2302</td>\n",
       "      <td>0.2691</td>\n",
       "      <td>0.2764</td>\n",
       "      <td>0.1616</td>\n",
       "      <td>0.2998</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   TEST.9  TEST.11  TEST.5  TEST.8  TEST.10  TEST.13  TEST.3  TEST.1  TEST.2  \\\n",
       "1  1.7733   0.1397  1.9420  0.7721   0.3296   0.7509  1.4407  1.1497  3.2036   \n",
       "2  0.4875   0.0846  0.2103  0.1855   0.3510   0.4165  0.3036  0.3107  0.1329   \n",
       "3  0.5832   2.3266  2.4897  1.1922   0.4258   1.2165  2.9190  1.7594  3.0148   \n",
       "4  0.3514   0.1188  0.1770  0.0979   0.0737   0.1280  1.1618  0.0345  0.1147   \n",
       "5  0.7182   0.3497  0.5963  0.1841   0.1702   0.2474  0.1502  0.2663  0.2369   \n",
       "\n",
       "   TEST.4   ...     TEST.20  TEST.17  TEST.18  TEST.22  TEST.16  TEST.23  \\\n",
       "1  2.3189   ...      3.0381   3.7487   0.1860   0.7623   0.8264   0.6403   \n",
       "2  1.2901   ...      0.1378   1.6411   0.0872   0.3245   0.2202   0.2515   \n",
       "3  0.8116   ...      1.2505   1.0424   1.9834   0.2148   2.2806   1.3489   \n",
       "4  0.1167   ...      0.2309   0.2140   0.1222   0.0707   0.0915   0.0646   \n",
       "5  0.2203   ...      0.3002   0.1965   0.1836   0.3171   0.3287   0.2302   \n",
       "\n",
       "   TEST.14  TEST.25  TEST.15  TEST.19  \n",
       "1   0.6729   0.8249   0.1181   0.7173  \n",
       "2   0.3038   0.3454   0.1068   0.1108  \n",
       "3   2.0071   1.2253   0.3881   0.6291  \n",
       "4   0.1553   0.1277   0.0622   0.0742  \n",
       "5   0.2691   0.2764   0.1616   0.2998  \n",
       "\n",
       "[5 rows x 25 columns]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "khan_train_df = pd.read_csv('./data/khan_train.csv', index_col=0)\n",
    "khan_test_df  = pd.read_csv('./data/khan_test.csv', index_col=0)\n",
    "display(khan_train_df.head())\n",
    "display(khan_test_df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hmm, I can't locate the response vectors..."
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}