SVM.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "hawaiian-audio",
   "metadata": {},
   "source": [
    "## SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "great-duplicate",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import sklearn\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.model_selection import train_test_split, StratifiedKFold, cross_val_score, learning_curve, ShuffleSplit, validation_curve\n",
    "from sklearn.metrics import plot_confusion_matrix\n",
    "import numpy as np\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.svm import SVC\n",
    "from imblearn.pipeline import Pipeline\n",
    "from category_encoders.one_hot import OneHotEncoder\n",
    "from sklearn.impute import SimpleImputer\n",
    "from imblearn.over_sampling import RandomOverSampler, SMOTE\n",
    "from mlxtend.evaluate import bias_variance_decomp\n",
    "from mlxtend.plotting import plot_decision_regions\n",
    "from sklearn.metrics import classification_report\n",
    "from imblearn.under_sampling import RandomUnderSampler\n",
    "\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.model_selection import RepeatedStratifiedKFold\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn import tree\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.metrics import average_precision_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "serious-airplane",
   "metadata": {},
   "outputs": [],
   "source": [
    "def my_learn_curve(title=\"some title\", \n",
    "                   estimator=None, \n",
    "                   cv = None ,\n",
    "                   X=None,\n",
    "                   y=None,\n",
    "                   label_suffix =  None,\n",
    "                   color_= [],\n",
    "                   shape='o-'):\n",
    "    plt.title(title)\n",
    "    plt.ylim(0.0, 1.01)\n",
    "    plt.xlabel(\"Training examples\")\n",
    "    plt.ylabel(\"Score\")\n",
    "\n",
    "    train_sizes, train_scores, test_scores, fit_times, _ = \\\n",
    "        learning_curve(estimator, X, y, cv=cv, n_jobs=4,\n",
    "                       train_sizes=np.linspace(.1, 1.0, 10),\n",
    "                       return_times=True)\n",
    "    train_scores_mean =  1.0 - np.mean(train_scores, axis=1)\n",
    "    train_scores_std = np.std(train_scores, axis=1)\n",
    "    test_scores_mean =  1.0 - np.mean(test_scores, axis=1)\n",
    "    test_scores_std = np.std(test_scores, axis=1)\n",
    "    fit_times_mean = np.mean(fit_times, axis=1)\n",
    "    fit_times_std = np.std(fit_times, axis=1)\n",
    "\n",
    "\n",
    "    # Plot learning curve\n",
    "    plt.grid()\n",
    "    plt.fill_between(train_sizes,  train_scores_mean - train_scores_std,\n",
    "                         train_scores_mean + train_scores_std, alpha=0.1,\n",
    "                         color=\"r\")\n",
    "    plt.fill_between(train_sizes,  test_scores_mean - test_scores_std,\n",
    "                         test_scores_mean + test_scores_std, alpha=0.1,\n",
    "                         color=\"g\")\n",
    "    plt.plot(train_sizes, train_scores_mean, shape, color=color_[0],\n",
    "                 label=\"Training Error - {}\".format(label_suffix))\n",
    "    plt.plot(train_sizes, test_scores_mean,shape, color=color_[1],\n",
    "                 label=\"Cross-validation Error - {}\".format(label_suffix))\n",
    "    plt.legend(loc=\"best\")\n",
    "    plt.ylabel(\"Error\")\n",
    "    return plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "sudden-convert",
   "metadata": {},
   "outputs": [],
   "source": [
    "SVC?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "colored-marketing",
   "metadata": {},
   "outputs": [],
   "source": [
    "parent_dir = '/mnt/c/users/17752/Documents/GeorgiaTech/machine_learning'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "popular-pasta",
   "metadata": {},
   "source": [
    "# Liver Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "severe-majority",
   "metadata": {},
   "outputs": [],
   "source": [
    "cross_val_score?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "extended-scale",
   "metadata": {},
   "outputs": [],
   "source": [
    "parent_dir = '/mnt/c/users/17752/Documents/GeorgiaTech/machine_learning'\n",
    "liver_df = pd.read_csv(f\"{parent_dir}/data/liver.csv\")\n",
    "liver_df = pd.read_csv(f\"{parent_dir}/data/liver.csv\")\n",
    "liver_df.drop_duplicates(inplace=True)\n",
    "impute_Albumin_and_Globulin_Ratio = liver_df[~liver_df.Albumin_and_Globulin_Ratio.isna()].Albumin_and_Globulin_Ratio.mean()\n",
    "filt_ =  liver_df.Albumin_and_Globulin_Ratio.isna()\n",
    "liver_df.loc[filt_, \"Albumin_and_Globulin_Ratio\" ]= impute_Albumin_and_Globulin_Ratio\n",
    "assert sum(liver_df.Albumin_and_Globulin_Ratio.isna()) == 0 \n",
    "# format response \n",
    "liver_df.loc[(liver_df.Dataset ==2), \"Dataset\" ] = -1\n",
    "liver_df.head(10)\n",
    "liver_df.loc[:,\"Gender\"] = pd.get_dummies(liver_df.Gender, prefix='gender').drop(\"gender_Male\", axis =1  ).gender_Female\n",
    "X = liver_df.loc[:, \"Age\":\"Albumin_and_Globulin_Ratio\"]\n",
    "y = liver_df[\"Dataset\"]\n",
    "X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, random_state=62)\n",
    "\n",
    "#('over', SMOTE(sampling_strategy=\"minority\")\n",
    "#inital shot \n",
    "liver_steps = [[('scaler', StandardScaler()),\n",
    "               (\"model\",  SVC(gamma='scale',kernel='rbf', cache_size=10000))],\n",
    "               [('scaler', StandardScaler()),\n",
    "               (\"model\",  SVC(gamma='scale',kernel='linear', cache_size=10000, probability=True)) ]]\n",
    "liver_pipeline_linear = Pipeline(steps=liver_steps[1])\n",
    "liver_pipeline_rbf =  Pipeline(steps=liver_steps[0])\n",
    "stratified_kfold = StratifiedKFold(n_splits = 10, random_state=42, shuffle=True)\n",
    "scores = cross_val_score(liver_pipeline_linear, X_train, y_train, scoring='recall', cv=stratified_kfold, n_jobs=3)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "familiar-dominant",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "arabic-given",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "2+2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "labeled-strain",
   "metadata": {},
   "source": [
    "### Base learning curve "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "obvious-cutting",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "my_learn_curve(title=\" Support Vector Machine - Liver\", \n",
    "               estimator=liver_pipeline_linear, \n",
    "               cv=stratified_kfold, \n",
    "               X=X_train,\n",
    "               y=y_train,\n",
    "               label_suffix=\"linear\", \n",
    "               color_=[\"orange\",\"blue\"],\n",
    "               shape = 'o-'  )\n",
    "\n",
    "# my_learn_curve(title=\" Support Vector Machine\", \n",
    "#                estimator=liver_pipeline_rbf, \n",
    "#                cv=stratified_kfold,\n",
    "#                X=X_train,\n",
    "#                y=y_train ,\n",
    "#                label_suffix=\"rbf\", \n",
    "#                color_=[\"orange\",\"blue\"], shape = 'x-' )\n",
    "\n",
    "\n",
    "plt.savefig(\"SVM_LIVER_BASE_learning_curve.pdf\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "disciplinary-apache",
   "metadata": {},
   "source": [
    "## average_precision_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "olive-cliff",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "aquatic-bahamas",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6973684210526315\n",
      "0.7719298245614035\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_train, liver_pipeline_linear.fit(X_train,y_train).predict(X_train)))\n",
    "print(average_precision_score( y_test, liver_pipeline_linear.fit(X_train,y_train).predict(X_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "thousand-memorabilia",
   "metadata": {},
   "source": [
    "###  Base Metrics "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "direct-priority",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.00      0.00      0.00       138\n",
      "           1       0.70      1.00      0.82       318\n",
      "\n",
      "    accuracy                           0.70       456\n",
      "   macro avg       0.35      0.50      0.41       456\n",
      "weighted avg       0.49      0.70      0.57       456\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.00      0.00      0.00        26\n",
      "           1       0.77      1.00      0.87        88\n",
      "\n",
      "    accuracy                           0.77       114\n",
      "   macro avg       0.39      0.50      0.44       114\n",
      "weighted avg       0.60      0.77      0.67       114\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/flufylobster/anaconda3/envs/ML39/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/home/flufylobster/anaconda3/envs/ML39/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1221: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "# print(grid_result_l.best_estimator_.get_params())\n",
    "print(classification_report(y_train, liver_pipeline_linear.fit(X_train,y_train).predict(X_train) ))\n",
    "print(classification_report(y_test, liver_pipeline_linear.fit(X_train,y_train).predict(X_test) ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "commercial-sponsorship",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(estimator=  liver_pipeline_linear.fit(X_train,y_train), X=X_train, y_true= y_train  , \n",
    "                      cmap=plt.cm.Blues,\n",
    "                       normalize=\"true\",\n",
    "                      display_labels = [\"Not Patient\" ,\"Patient\"]\n",
    "                        )\n",
    "\n",
    "plt.title(\" SVM - Confusion Matrix Liver\")\n",
    "\n",
    "plt.savefig(\"SVM_LIVER_BASE_CONFUSION_MAT.PDF\")\n",
    "### Overfit "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "temporal-senate",
   "metadata": {},
   "source": [
    "## Validation/ complexity  Curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "municipal-sarah",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "declared-fetish",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAv2UlEQVR4nO3deZwcdZ3/8denu+dIMjkgCQQSJEHDnZMh4b6XICjIKj9IkCP+VkAEFBYE3VVZXVx1dUFcBBE5BTmWQ5T8BFmIyJ0EAxIIIWKA4QyJuSaZo7s/vz+quqe6p2emZzI9k0y9n49HP6rrW9+u/lYn8/lUfetbVebuiIhIfCX6uwEiItK/lAhERGJOiUBEJOaUCEREYk6JQEQk5pQIRERiTolAesTM3Mw+Eb6/zsy+WU7dHnzPqWb2SE/bKX3DzD5mZhvMLNnfbZHuUyKIKTN72My+U6L8BDN738xS5a7L3c9x9+/2QpvGh0kj/93ufru7H7256+7g+4aZ2VVm9lYYxJaH86Mq8X3dbNtBZva0ma01s9Vm9pSZ7Wtm+5tZo5kNLfGZP5vZeZHf8YWi5aPMrMXMVmxGu+ab2T8Vl7v7W+5e5+6Znq5b+o8SQXzdDJxmZlZUfhpwu7un+75JfcfMqoH/BfYCjgGGAQcAq4AZPVhf2YmzjHUNA34H/BTYFhgL/BvQ7O7PAA3AZ4s+szewJ/DrSPGQsDxnDvC33mpnX+jN31U6pkQQXw8QBJmDcwVmtg3wKeBWM5thZs+Y2Roze8/M/jsMnu2Y2c1m9u+R+UvCz7xrZl8oqntcuOe6zszeNrPLI4ufCKdrwj30/c3sTDN7MvL5A8xsQbinvMDMDogsm29m3w33nteb2SOd7N2fDnwMONHdX3H3rLt/6O7fdfd54foKurSi22lmh5lZg5ldambvAzeZ2atm9qlI/ZSZfWRm08P5/cK9/DVm9qKZHdZB23YFcPdfu3vG3Te5+yPu/lK4/Jaw/cXb85C7r4qU3QacUVTn1g6+c7NEj+bM7BQzW1i0/EIzezB8X2NmPwqPxD4IuxYHhcva/a6VaK8UUiKIKXffBNxNYUD5P8BSd38RyAAXAqOA/YEjgXO7Wq+ZHQNcDPwDMBE4qqhKY/idI4DjgC+Z2WfCZYeE0xFhN8MzReveFngIuBoYCfwX8JCZjYxUmwPMBbYDqsO2lHIU8Ht339DVNnViDEEy3Rk4i2BvfHZk+SzgI3d/wczGhm3/9/AzFwP3mtnoEutdBmTM7BYz+2SYoKNuAw42s48BmFmCYLuLg/yvgFPMLGlmewBDged6vrllexDYzcwmRsrmAHeE739AkOymAp8gOOL5VqRu8e8qFaZEEG+3ACfl9sYIAvQtAO6+yN2fdfe0u68Afg4cWsY6/w9wk7u/7O6NwOXRhe4+393/Eu6Bv0QQPMtZLwSJ43V3vy1s16+BpcCnI3VucvdlkUQ3tYN1jQTeK/N7O5IFvu3uzeH33QEcb2aDw+XR4Pd5YJ67zwu3/Q/AQuDY4pW6+zrgIMCBXwArzexBM9s+XP428MdwnRAk6VqCRBPVALxGkPTOoEJHA8XcfSPwG8KkGCaE3YEHw67ILwIXuvtqd18PfA84JbKK4t9VKkyJIMbc/UlgJXCCme0C7EsYuMxsVzP7nQUnjtcR/LGWcxJ1R+DtyPyb0YVmNtPMHjezlWa2FjinzPXm1v1mUdmbBHuUOe9H3m8E6jpY1ypghzK/tyMr3b0pN+Puy4FXgU+HyeB42hLBzgRJd03uRRDsS7bB3V919zPdfRywN8G2XxWpEu0eOg24w91bS6zqVuBMgqD8q842xsy+EXbJbTCz6zqrW4Y7aDs6mgM8ECaI0cBgYFHkd/h9WJ5T8LtK5SkRyK0EAeU04BF3/yAsv5Zgb3uiuw8DvgEUn1gu5T1gp8j8x4qW30HQdbCTuw8Hroust6tb4b5LEFCjPga8U0a7ij0KzDKzIZ3U2UgQtHLGFC0v1d5c99AJwCthcoAgOd7m7iMiryHu/v2uGuruSwlO7kdP/N4HjDWzw4F/pOO9/XsJjqTecPfiJFr8Pd8Lu+Tq3P2crtrVhUeAUWY2leD3yCXEj4BNwF6R32G4u0cTtm6J3MeUCORWgq6DLxJ2C4WGAuuADWa2O/ClMtd3N3Cmme0Z7hV/u2j5UGC1uzeZ2QyCvcWclQTdArt0sO55wK5mNic8KXkywUiZ35XZtqjbCILzvWa2u5klzGxkuFec665ZDMwJ+9iPobwurDuBowl+rzsi5b8iOFKYFa6vNjwxOq54BWF7/jm3zMx2Igimz+bqhN1u/0NwMvVNd19YvJ5IvSOAdkM+N0MqbH/uVVXie9Nh+/6ToL//D2F5lqC760oz2y7cvrFmNqsX2yfdpEQQc2H//9PAEII99ZyLCYL0eoI/3LvKXN//I+jCeAxYHk6jzgW+Y2brCU4Q3h357EbgCuCpsNtgv6J1ryIY1fTPBF07XwM+5e4fldO2onU1EyTApQRBah3wPEE3Ve6E6lcIzj+sAU4lGGnV1XrfA54hGIp6V6T8bYKjhG8QJLy3gUso/Te4HpgJPGdmjQQJ4GWC7Y66heAIqdO+f3df6O5/7art3XAtwV597tXRyJ47CH7je4qGI19K8H/j2bDb8VFgt15sn3ST6cE0IiLxpiMCEZGYUyIQEYk5JQIRkZhTIhARibmt7oZOo0aN8vHjx/d3M0REtiqLFi36yN1L3dJk60sE48ePZ+HCkkOmRUSkA2bW4QWF6hoSEYk5JQIRkZhTIhARibmt7hyByEDW2tpKQ0MDTU26+ab0TG1tLePGjaOqqt0toDqkRCCyBWloaGDo0KGMHz8ea/cUUZHOuTurVq2ioaGBCRMmlP05dQ2JbEGampoYOXKkkoD0iJkxcuTIbh9RKhGIbGGUBGRz9OT/jxKBiEjMKRGISN6qVauYOnUqU6dOZcyYMYwdOzY/39LS0ulnFy5cyAUXXNDldxxwwAG91VzpJTpZLCJ5I0eOZPHixQBcfvnl1NXVcfHFF+eXp9NpUqnSYaO+vp76+vouv+Ppp5/ulbb2ts62baDTEYGIdOrMM8/koosu4vDDD+fSSy/l+eef54ADDmDatGkccMABvPbaawDMnz+fT33qU0CQRL7whS9w2GGHscsuu3D11Vfn11dXV5evf9hhh/G5z32O3XffnVNPPZXcg7LmzZvH7rvvzkEHHcQFF1yQX2/UkiVLmDFjBlOnTmXy5Mm8/vrrANx6661MnjyZKVOmcNpppwHw5ptvcuSRRzJ58mSOPPJI3nrrrZLb9te//pVjjjmGffbZh4MPPpilS5dW6FfdssQz/YlsDX5coZPG/9z9pxIuW7aMRx99lGQyybp163jiiSdIpVI8+uijfOMb3+Dee+9t95mlS5fy+OOPs379enbbbTe+9KUvtRvb/uc//5klS5aw4447cuCBB/LUU09RX1/P2WefzRNPPMGECROYPXt2yTZdd911fOUrX+HUU0+lpaWFTCbDkiVLuOKKK3jqqacYNWoUq1evBuC8887j9NNP54wzzuDGG2/kggsu4IEHHmi3bUceeSTXXXcdEydO5LnnnuPcc8/lsceKn7Y68CgRiEiXTjrpJJLJJABr167ljDPO4PXXX8fMaG1tLfmZ4447jpqaGmpqathuu+344IMPGDduXEGdGTNm5MumTp3KihUrqKurY5dddsmPg589ezbXX399u/Xvv//+XHHFFTQ0NPCP//iPTJw4kccee4zPfe5zjBo1CoBtt90WgGeeeYb77rsPgNNOO42vfe1r7bZtw4YNPP3005x00kn5Zc3NzT36vbY2SgQiW6oe7LlXypAhQ/Lvv/nNb3L44Ydz//33s2LFCg477LCSn6mpqcm/TyaTpNPpsuqU+xz1OXPmMHPmTB566CFmzZrFDTfcgLuXNXwyWie3bdlslhEjRuTPkcSJzhGISLesXbuWsWPHAnDzzTf3+vp333133njjDVasWAHAXXfdVbLeG2+8wS677MIFF1zA8ccfz0svvcSRRx7J3XffzapVqwDyXUMHHHAAd955JwC33347Bx10ULv1DRs2jAkTJnDPPfcAwVW6L774Ym9v3hapYonAzG40sw/N7OUOlpuZXW1my83sJTObXqm2iEjv+drXvsbXv/51DjzwQDKZTK+vf9CgQfzsZz/jmGOO4aCDDmL77bdn+PDh7erddddd7L333kydOpWlS5dy+umns9dee/Ev//IvHHrooUyZMoWLLroIgKuvvpqbbrqJyZMnc9ttt/GTn/yk5Hfffvvt/PKXv2TKlCnstdde/OY3v+n17dsSWbmHYd1esdkhwAbgVnffu8TyY4HzgWOBmcBP3H1mV+utr693PZhGBqpXX32VPfbYo7+b0e82bNhAXV0d7s6Xv/xlJk6cyIUXXtjfzdpqlPp/ZGaL3L3k+N6KHRG4+xPA6k6qnECQJNzdnwVGmNkOlWqPiGw9fvGLXzB16lT22msv1q5dy9lnn93fTRrQ+vNk8Vjg7ch8Q1j2XnFFMzsLOAvgYx/7WJ80TkT6z4UXXqgjgD7UnyeLS53aL9lP5e7Xu3u9u9ePHl3y2csiItJD/ZkIGoCdIvPjgHf7qS0iIrHVn4ngQeD0cPTQfsBad2/XLSQiIpVVsXMEZvZr4DBglJk1AN8GqgDc/TpgHsGIoeXARmBupdoiIiIdq+SoodnuvoO7V7n7OHf/pbtfFyYBwtFCX3b3j7v7JHfXmFCRLcD777/PKaecwsc//nH23HNPjj32WJYtW9bfzWrn5ptv5rzzzgOC+w7deuut7eqsWLGCvfduN3q9XZ077rgjP1/u7bQHEt1iQkTy3J0TTzyRM844I38l7uLFi/nggw/Ydddd8/UymUz+3kNbgnPOOafHn80lgjlz5gDl3067r1XyN9ctJkQk7/HHH6eqqqogsE6dOpWDDz6Y+fPnc/jhhzNnzhwmTZpEU1MTc+fOZdKkSUybNo3HH38cKH176MbGRo477jimTJnC3nvv3e62EdlslvHjx7NmzZp82Sc+8Qk++OADfvvb3zJz5kymTZvGUUcdxQcffNCu3Zdffjk/+tGPAFi0aBFTpkxh//3355prrsnXWbFiBQcffDDTp09n+vTp+eciXHbZZfzpT39i6tSpXHnllQW30169ejWf+cxnmDx5Mvvttx8vvfRS/vs6us12TiaT4cwzz2Tvvfdm0qRJXHnllQAsX76co446iilTpjB9+nT++te/4u5ccskl+bq536f4N89kMlxyySXsu+++TJ48mZ///Ofd+wfugI4IRLZQZv9WkfW6f7vDZS+//DL77LNPh8uff/55Xn75ZSZMmMCPf/xjAP7yl7+wdOlSjj76aJYtW1by9tDz5s1jxx135KGHHgKC+xVFJRIJTjjhBO6//37mzp3Lc889x/jx49l+++056KCDePbZZzEzbrjhBn74wx/mv7uUuXPn8tOf/pRDDz2USy65JF++3Xbb8Yc//IHa2lpef/11Zs+ezcKFC/n+97/Pj370I373u98BQfDN+fa3v820adN44IEHeOyxxzj99NPzN6Xr6jbbixcv5p133uHll4O77OSS3Kmnnspll13GiSeeSFNTE9lslvvuu4/Fixfz4osv8tFHH7HvvvtyyCGHtPvNr7/+eoYPH86CBQtobm7mwAMP5Oijj87fqbWndEQgImWbMWNGPug8+eST+Qe/7L777uy8884sW7aM/fffn+9973v84Ac/4M0332TQoEFMmjSJRx99lEsvvZQ//elPJe8ddPLJJ+f3hO+8805OPvlkABoaGpg1axaTJk3iP//zP1myZEmH7Vu7di1r1qzh0EMPBci3D6C1tZUvfvGLTJo0iZNOOolXXnmly+2NbuMRRxzBqlWr8kksd5vtUaNG5W+zHbXLLrvwxhtvcP755/P73/+eYcOGsX79et555x1OPPFEAGpraxk8eDBPPvkks2fPJplMsv3223PooYeyYMGCdr/5I488wq233srUqVOZOXMmq1atyj+QZ3PoiEBkC9XZnnul7LXXXvzP//xPh8ujt6Pu6D5lpW4PfcQRR7Bo0SLmzZvH17/+dY4++mhmzZqVv3XEd77zHT796U+zfPlyVq5cyQMPPMC//uu/AnD++edz0UUXcfzxxzN//nwuv/zyDtvX2W2or7zySrbffntefPFFstkstbW1Xf0cJbcxt/6ubrO9zTbb8OKLL/Lwww9zzTXXcPfdd3PVVVeV/T05xb/5T3/6U2bNmtVl27tDRwQiknfEEUfQ3NzML37xi3zZggUL+OMf/9iu7iGHHMLtt98OBE/5euutt9htt91K3h763XffZfDgwXz+85/n4osv5oUXXmDmzJksXryYxYsXc/zxx2NmnHjiiVx00UXssccejBw5Eii87fUtt9zSaftHjBjB8OHDefLJJwHy7cutZ4cddiCRSHDbbbfl75w6dOhQ1q9fX3J90W2cP38+o0aNYtiwYWX9lh999BHZbJbPfvazfPe73+WFF15g2LBhjBs3Lv90tObmZjZu3MghhxzCXXfdRSaTYeXKlTzxxBPMmDGj3TpnzZrFtddem38Y0LJly2hsbCyrPZ3REYGI5JkZ999/P1/96lf5/ve/T21tLePHj+eqq67inXfeKah77rnncs455zBp0iRSqRQ333wzNTU13HXXXfzqV7+iqqqKMWPG8K1vfYsFCxZwySWXkEgkqKqq4tprry35/SeffDL77rtvwXMOLr/8ck466STGjh3Lfvvtx9/+9rdOt+Gmm27iC1/4AoMHDy7Ycz733HP57Gc/yz333MPhhx+e39OePHkyqVSKKVOmcOaZZzJt2rSC7547dy6TJ09m8ODBXSaiqHfeeYe5c+eSzWYB+I//+A8AbrvtNs4++2y+9a1vUVVVxT333MOJJ57IM888w5QpUzAzfvjDHzJmzJh2z0z+p3/6J1asWMH06dNxd0aPHp1PKpujYrehrhTdhloGMt2GWnrDFnMbahER2TooEYiIxJwSgcgWZmvrrpUtS0/+/ygRiGxBamtrWbVqlZKB9Ii7s2rVqrKGxkZp1JDIFmTcuHE0NDSwcuXK/m6KbKVqa2sZN25ctz6jRCCyBamqqtrs2wWIdJe6hkREYk6JQEQk5pQIRERiTolARCTmlAhERGJOiUBEJOaUCEREYk6JQEQk5pQIRERiTolARCTmlAhERGJOiUBEJOaUCEREYk6JQEQk5pQIRERiTolARCTmlAhERGKuoonAzI4xs9fMbLmZXVZi+XAz+62ZvWhmS8xsbiXbIyIi7VUsEZhZErgG+CSwJzDbzPYsqvZl4BV3nwIcBvzYzKor1SYREWmvkkcEM4Dl7v6Gu7cAdwInFNVxYKiZGVAHrAbSFWyTiIgUqWQiGAu8HZlvCMui/hvYA3gX+AvwFXfPFq/IzM4ys4VmtnDlypWVaq+ISCxVMhFYiTIvmp8FLAZ2BKYC/21mw9p9yP16d6939/rRo0f3djtFRGKtkomgAdgpMj+OYM8/ai5wnweWA38Ddq9gm0REpEglE8ECYKKZTQhPAJ8CPFhU5y3gSAAz2x7YDXijgm0SEZEiqUqt2N3TZnYe8DCQBG509yVmdk64/Drgu8DNZvYXgq6kS939o0q1SURkq+AOePupJSGR7PWvq1giAHD3ecC8orLrIu/fBY6uZBtEJGY6CqLF04KybJn1NqNO9Lu6Wp5thUxL8Mq2QLolKKuqhZ1nQaqmV3+yiiYCEYkRzwXUbCTYdTJfVp1I4OxxQAY8A5nWIJhmW8HTYZBNB4E2kw6XhdNMK3hrYVm2NazXEllXum0d0XrZdCfT6CtsS7TMMx3/xiP3gnGHKRGISBmiQbmcYEu2i7Lo+joqK9qzzmYg0wzpTZDZBOnwfXoTZJrCZU3BK/c+0xxZ1hwJsLkgWRQwiwNoriy/LBN8pv2o9C1bIgWJqsgrBWYwqDKjJpUIRCqhZJdDNMBGuwiK6pSz55z/fBdBObdnHF1HLvjmg3JLGKijQXlT+6CcKX61FE6zLW3dGZmwK6PdiPF+lEiBpSBZVThNpCKBNxlMLRmZT0WCcbR+5DPF5ZYsmqbCeuErWQUk2sqiUxJg4TgeSwQJAAumQ3aoyE+jRCDx497xHmZxd0RHwbrsgF6ib7ggOBN8d35vuDkSiHMBOxeQm9oCcropDLhNYR9yJDhnm4Oui0xz2J1RNN+XwTlRBalaSNZAsjbo0kjWQrK6xLQ6rFcNiWpIVRcGWMsF3ERbYE0kgVywDsuxcD4XeHOB1dr+zXJdRjj5S56iATf3PvrKleeDtBEMvIzUt0RYPVHisyWCe+4zWOHncuu26PJE8NtUDen1fyYlAtn6tet3LQrwJctyXQaZ4H02HQbblrbgmy0OtC1twTcfWKN7wC1tQTfbGnZrtBb2Tef2lDORZd7Hd1XpMDjXhOXVkKgJy2va5vPBujrYo02EATsZdl8kqwu7MpJVgBV1M4WBNxfg8sE0UVhmRWW5QBkNutEgm59S+BkoXc+s/boKgnmp5cXtKQrSpeqU/Fxku4u/syAR9B0lAtmy5IJyl4G8uCwTCeqt0LIWmtZA89/D6TpoWQPNa6F1A7RsCKfroXV9MO23fmQr2jOuaQvU0YCdrA7mE0V7z8maMLhXR4Jz+EqkCoN1ItxjbndkEw3SifbBqrisuJ51UkZRWakXHZRH9+aLg3OH5cWBuKtAL0oE0neymcL+5Fx3RvSEYC6Y5072ZVqhZR00/T0I6s1rgmDevCYI7rkgng/umxHUE6kwqFYHe8OJ6sgecuR9QQCOBt3onnJVWyDO7ymn2pbl9pgtRb47oV3ffzRY59qYLH8vOlpWvNdZag+2nIDcK/UUfLc0SgTSe9wjJxA7OZnYsgEa34XG92HjB7Dp78EefMu6wqCe21vvSVCvGgI1I6BmeOG0eljwvnooVNW1TVODw/5mp13/f7tzAN3t5uiNvehoWaLwpGJvBe1cH7bEjhKBdE++n7y5fdBPN0HTamh8L3y9D5tWQuOH0PRRsKxpNbQ2du87OwzqYUCvGQpVQ6G6LgjqVUOCoJYNhw3mjzLSwbnB6OiNgveRwNnVHnRBAO8o8FoH5Z11fxSXJyPfufXwyFFM7m2uLHqAU1xW6nNBLrUOp7L5lAikUDZdeq8+3RQE9g3vBnvzG96DjR/Cpg9h40poWhUE+Uxz19+RqILB2wWvQaNh0MhgT716aNu0akjwqq4DEoXjwz0D2WwYISIBvXgESXJQ2/C/giF6RUP8Cob3JXA3sm44CbL594Vl7rSVZx13J5vxoDzrbWXZtrLS8xmy2XRBWTRQdhbnOguCPf1c8Wd7HsRLJYLCaTjXyTra2pML/GD5g6y28qAskbBOE0awvPOk0lmdZNJIpRIkkwmSSSuYbu2UCOKi5JDJcPRK6wZY3xC8NrwLG99v67bZ+FGwV9+0mk6veMxJDY4E+VFBkK8dBbXbQO22wTQ1GMhGzglk27o78uOuE5HgHQ3kCQrGZBeM6U6R9QQZT5LOJshkg/cZD96nMxa8z2TJZDw/TaezZDLZMICnwwCdG/nZFsAhF+QB2oJ6W532n8kF9nI+k6tfjr5IAlBeIiis31ZW6rtyZdEBPh3Vz81G25AbjZv7/XKfKw7cnSWMntRJJIKyVCpBIhEkjGQyEU47ThBBubVbVly/vykRbK2iQx6b1oQnT9cE/ezNawtPpLasbxsl09oI6UZo3RjU37gymJYztrx6OAweHQnwI9uCe802QZdNsiZol1m4h56CZHTvOxkJ5KnIHnvwPpsL4p4sDOLZBJlMgkw2LPMEmWwuiHtBcM9mnUwmQybTGr53stksmQzh8uB9NhvUTac9/97d2+1lAu32FKN7n8V7orn5VCrRrc9Ev6ezfNBZsigzj5S9ztLBvP2yUmV9oS05eImEUXravTptybzt/5bnj/rS6eD8VZAcEiSThNO2ZJF7XzwfTSQdJYjihFJTk6Kurvef5hubRLBx3TqWvvI+jRtb+7spmLeSyGwkmWkkmW4kkd5AIrOBRHoDycwGEun1JDONJDIbguWZxqB+diOJzKZgmt1EMruJhPd8e2pSaWpSGRyjNbkNramRtKS2pTW1La2pEbQkt6E1MZzW1AhaE8PxZC1YAst1vWRSsCmBNUW6Y0hiydzeenCFpiWjV2sGScFSVUAynyQy4R9Y7o8tnc7m/9gymXQYvAuDejAtDOpAyT/C6B9dTU1V5A+zcHkuIPe3zuPpltHGLUF0z72/uBcmh9xOSe59Nuu0tGTbLcslldzOR/D/tTCRJBKFRyF1ddVMmrQ9gwdX9eo2xCYRXPz5K7j2t4P7uxllMGBY+Kq8QTVQNyTJ0CFJ6gYnGVoXvoZUUVeXYujQKobWha+hNdQNrSGVinbNJLGCk64JLBl27eS2qOBvNANkgcIEVm4gT6WMZLJ0IE8mTScPpc8FR389/39XKpFki3aMWlqyrF/fTCqVzB+F9KbYJILamgRDa8o4kdlnCjpHw44ZK1oWTL3gEvjC5WX2KJcsbW7JsqnZ2dScYeXqMvr/Q0OHVjN8eC0jRtSE0+A1fHhN5H1ueQ11ddX5ftDiHojc3lAuqA/EQJ77Q29tzdDamqWlJUM6HUyj71tbgyOcIUOqGTq0mmHDgt8uler/PmSpnHITSUtL+X+j3RWbRPDDX1/ByU+9wt+WvkWf3mulIwWjXHIXCVnbNM/aXvmOWIrqFiaVws8UlUcCrbuzYUMLTU1pmprSbNzYysaNrTQ2trB+fQvr1jWzZk1T+Arer13bxPr1wfKGhjI3NWEMG1ZTkCiir5qarh+00d2TnD2RTmdpbc3mA3ZrayYfoIvfd1TWUZDvad89wODBVQwdWs3QoTUF02HDatqVDxtWOD9oUGpAJlfpXbFJBKlUglHjdmTQNpW5jevWyN3zASv3am7O0NoaTM2gujqZf9XUJEkmjebmDGvXRpNEU4n5tgQSTShvvrm2vze7X+ROBFZXJ6mqSlJVlaCqKkl1dSKcD94nEkZjY2uYbJtZv74ln6A/+KCb11+E31sqQUSPOAoTSlB38OCqXjlf0htJqHCkTls3ofSe2CQCgI9/fNv+bsIWJ5PJ0tycyR8VNDen8++LE8OaNc3hHnCGVCrByJGDGDOmjpqaXKJIlezGSKezRUcXha/W1s77PLsaUlnO3nY568gF6urqZCRoJ/KBOxrES9XprH5Phwi6O42NrWzYEByhrVvXnH8fTRaF5W1lwb9b8DsPNB2NzOkocQR1ciNxgve5805to34K6+XWX1WVoLY2VfJVU5Msq3xLGCbakVglAmkvmUwweHCi5CiE1ta2BBEkidx8Ky0t2XyCaGpKs3Ztc76PO5cY2o4kUgwfXsO22w7qhy3cupkFI0Xq6qoZM6au259vacnkE8P69c2sW9dSMN9ReWPj5o+uK/eaiM7XQbtROJlMsN7cKLKudiS2FNFkUlMTTRrJogRSOuFs2NDCTjsNZ9q0Mb3eNiUC6VCuy2Lo0MLH4rl7wVFE9EiiuTlDS0ualpZsGIRaWbVqEy0tGZJJyyeGqqoENTWp/F60+rEro7o6yciRgxk5cmsYMVe+zkbYFL8vNWyzcHhycf3iocpBsin1/734Vfx3EX0F55KC82s9tcceo/jiF6f34i8ZUCKQbjOz/F5KsXQ62+6PJPfH0dbVFJQ1Nm7KdzUVHr63Hzra0UU4W9LYf+k70W6grYF7uckk02GSWb16EzvvPKIi7VMikF6VSiVIpaoZMqTw6sfcielSRxLNzekOr/TN7a3lriAudcEOUDI5dN5X3P1rD9puF+EF85t7VWuuLHf7iWTSCk4o64hp62dm+a7SYcN69uD5t95ay7bbVubITolA+oRZcHl8TU2q3R9Cbpx9NOAX3weos+WlrjouvpVEa2umw2QDFCSE4nva5Mra7pMTvf1E+/vVROu0n3ZUJ0hMwe0lnI0bW/PDUtPpbJhgS5/ADm5loUQhPadEIP0ud0FNKpWgpgc7S9GjhJ4kk7YbzjnRWxZ0dUOy7t7Zsty6mUzu6Cmdn6bTnu9Ga23NsnFja/46hVyiiI5eiiYNJQrpihKBbPWC7p8kVT28/Uq0q6nUDeGie/H9Idqt1na9RzpSlg4vZsuSTgfLGxtb80NzM5lsp8Nht5Z+dqkcJQKJvdzNvrZU0W61UoJ70ZRKEG3zQTdTltbWIGmsX99KOt1ES0uGbNZLHk1ET8IX3120s7uNVqquVI4SgchWLpGIjuJq37eWSxQdJYnC22RkaGnJsm5dC9ngIQolHxqT0/7hNIXTcK7kejp7tkHxfHcSQ2cJp6uyrr4repQYrRudLz6CLDwvVHpZOeurxM3mcpQIRAa4wkTRXiaTJXqLkVySaBsdRTgt/yll3anT2WMti98X1yunvHTd0iso/fmO2t9Wnnvl1l040KD99kTrB9Pih+60/0xtbYpUyjpMYJtDiUAk5pLJBIMGJRg0qHfvcd8bupcEyqvb0RXPXdUtNfw3V97R8ODeXl5VldCDaUQkXsrt5gmXVLQtA1lFhwuY2TFm9pqZLTezyzqoc5iZLTazJWb2x0q2R0RE2qvYEYGZJYFrgH8AGoAFZvagu78SqTMC+BlwjLu/ZWbbVao9IiJSWtlHBGY2yMx268a6ZwDL3f0Nd28B7gROKKozB7jP3d8CcPcPu7F+ERHpBWUlAjP7NLAY+H04P9XMHuziY2OBtyPzDWFZ1K7ANmY238wWmdnpHXz/WWa20MwWrly5spwmi4hImco9IricYA9/DYC7LwbGd/GZUmduis/Lp4B9gOOAWcA3zWzXdh9yv97d6929fvRoPWFMRKQ3lXuOIO3ua7t5tV8DsFNkfhzwbok6H7l7I9BoZk8AU4Bl3fkiERHpuXKPCF42szlA0swmmtlPgae7+MwCYKKZTTCzauAUoLg76TfAwWaWMrPBwEzg1W60X0RENlO5ieB8YC+gGbgDWAt8tbMPuHsaOA94mCC43+3uS8zsHDM7J6zzKsF5h5eA54Eb3P3lHmyHiIj0kHX1XNFwGOjD7n5U3zSpc/X19b5w4cL+boaIyFbFzBa5e32pZV0eEbh7BthoZsN7vWUiItLvyj1Z3AT8xcz+ADTmCt39goq0SkRE+ky5ieCh8CUiIgNMWYnA3W8JR/7kxvi/5u6tlWuWiIj0lbISgZkdBtwCrCC4UGwnMzvD3Z+oWMtERKRPlNs19GPgaHd/DSC8+vfXBFcFi4jIVqzc6wiqckkAwN2XAVveUyxERKTbyj0iWGhmvwRuC+dPBRZVpkkiItKXyk0EXwK+DFxAcI7gCYLnCIiIyFau3ESQAn7i7v8F+auNayrWKhER6TPlniP4X2BQZH4Q8GjvN0dERPpauYmg1t035GbC94Mr0yQREelL5SaCRjObnpsxs3pgU2WaJCIifanccwRfBe4xs3cJnjK2I3BypRolIiJ9p9MjAjPb18zGuPsCYHfgLiBN8AyBv/VB+0REpMK66hr6OdASvt8f+AZwDfB34PoKtktERPpIV11DSXdfHb4/Gbje3e8F7jWzxRVtmYiI9ImujgiSZpZLFkcCj0WWlXt+QUREtmBdBfNfA380s48IRgn9CcDMPkHw3GIREdnKdZoI3P0KM/tfYAfgEW97wHGC4IH2IiKyleuye8fdny1RtqwyzRERkb5W7gVlIiIyQCkRiIjEnBKBiEjMKRGIiMScEoGISMwpEYiIxJwSgYhIzCkRiIjEnBKBiEjMKRGIiMRcRROBmR1jZq+Z2XIzu6yTevuaWcbMPlfJ9oiISHsVSwRmliR4iM0ngT2B2Wa2Zwf1fgA8XKm2iIhIxyp5RDADWO7ub7h7C3AncEKJeucD9wIfVrAtIiLSgUomgrHA25H5hrAsz8zGAicC13W2IjM7y8wWmtnClStX9npDRUTirJKJwEqUedH8VcCl7p7pbEXufr2717t7/ejRo3urfSIiQmUfN9kA7BSZHwe8W1SnHrjTzABGAceaWdrdH6hgu0REJKKSiWABMNHMJgDvAKcAc6IV3H1C7r2Z3Qz8TklARKRvVSwRuHvazM4jGA2UBG509yVmdk64vNPzAiIi0jcqeUSAu88D5hWVlUwA7n5mJdsiIiKl6cpiEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJuYomAjM7xsxeM7PlZnZZieWnmtlL4etpM5tSyfaIiEh7FUsEZpYErgE+CewJzDazPYuq/Q041N0nA98Frq9Ue0REpLRKHhHMAJa7+xvu3gLcCZwQreDuT7v738PZZ4FxFWyPiIiUUMlEMBZ4OzLfEJZ15P8C/6/UAjM7y8wWmtnClStX9mITRUSkkonASpR5yYpmhxMkgktLLXf369293t3rR48e3YtNFBGRVAXX3QDsFJkfB7xbXMnMJgM3AJ9091UVbI+IiJRQySOCBcBEM5tgZtXAKcCD0Qpm9jHgPuA0d19WwbaIiEgHKnZE4O5pMzsPeBhIAje6+xIzOydcfh3wLWAk8DMzA0i7e32l2iQiIu2Ze8lu+y1WfX29L1y4sL+bISKyVTGzRR3taOvKYhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQIRkZhTIhARiTklAhGRmKtoIjCzY8zsNTNbbmaXlVhuZnZ1uPwlM5teyfaIiEh7FUsEZpYErgE+CewJzDazPYuqfRKYGL7OAq6tVHtERKS0Sh4RzACWu/sb7t4C3AmcUFTnBOBWDzwLjDCzHSrYJhERKZKq4LrHAm9H5huAmWXUGQu8F61kZmcRHDEAbDCz17r47lHAR91t8ACg7Y6fuG67trv7du5oQSUTgZUo8x7Uwd2vB64v+4vNFrp7fbn1Bwptd/zEddu13b2rkl1DDcBOkflxwLs9qCMiIhVUyUSwAJhoZhPMrBo4BXiwqM6DwOnh6KH9gLXu/l7xikREpHIq1jXk7mkzOw94GEgCN7r7EjM7J1x+HTAPOBZYDmwE5vbS15fdjTTAaLvjJ67bru3uReberkteRERiRFcWi4jEnBKBiEjMDbhE0NVtLQYiM7vRzD40s5f7uy19ycx2MrPHzexVM1tiZl/p7zb1BTOrNbPnzezFcLv/rb/b1JfMLGlmfzaz3/V3W/qKma0ws7+Y2WIzW9jr6x9I5wjC21osA/6BYGjqAmC2u7/Srw2rMDM7BNhAcJX23v3dnr4SXoW+g7u/YGZDgUXAZ2Lw723AEHffYGZVwJPAV8Kr8wc8M7sIqAeGufun+rs9fcHMVgD17l6Ri+gG2hFBObe1GHDc/QlgdX+3o6+5+3vu/kL4fj3wKsGV6QNaeEuWDeFsVfgaOHt0nTCzccBxwA393ZaBZKAlgo5uWSEDnJmNB6YBz/VzU/pE2D2yGPgQ+IO7x2K7gauArwHZfm5HX3PgETNbFN5yp1cNtERQ1i0rZGAxszrgXuCr7r6uv9vTF9w94+5TCa7Gn2FmA75L0Mw+BXzo7ov6uy394EB3n05wx+Yvh93BvWagJQLdsiJmwj7ye4Hb3f2+/m5PX3P3NcB84Jj+bUmfOBA4PuwvvxM4wsx+1b9N6hvu/m44/RC4n6AbvNcMtERQzm0tZIAIT5r+EnjV3f+rv9vTV8xstJmNCN8PAo4ClvZro/qAu3/d3ce5+3iCv+3H3P3z/dysijOzIeFgCMxsCHA00KsjBAdUInD3NJC7rcWrwN3uvqR/W1V5ZvZr4BlgNzNrMLP/299t6iMHAqcR7BkuDl/H9nej+sAOwONm9hLBzs8f3D02QyljaHvgSTN7EXgeeMjdf9+bXzCgho+KiEj3DagjAhER6T4lAhGRmFMiEBGJOSUCEZGYUyIQEYk5JQKRXmBmY8zsTjP7q5m9YmbzzGzX/m6XSDmUCEQ2U3hh2/3AfHf/uLvvCXyDYPy3yBavYs8sFomRw4HW8DncALj74v5rjkj36IhAZPPtTfAsBJGtkhKBiEjMKRGIbL4lwD793QiRnlIiENl8jwE1ZvbFXIGZ7Wtmh/Zjm0TKppvOifQCM9uR4OlZ+wBNwAqCB+W83o/NEimLEoGISMypa0hEJOaUCEREYk6JQEQk5pQIRERiTolARCTmlAhERGJOiUBEJOb+P5nN0IcFcLelAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# cv = ShuffleSplit(n_splits=10, test_size=0.1, random_state=0)\n",
    "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=1, random_state=45)\n",
    "train_scores, valid_scores = validation_curve(liver_pipeline_rbf,\n",
    "                                              X_train,\n",
    "                                              y_train, \n",
    "                                              param_name =  \"model__C\",\n",
    "                                            param_range = np.linspace(0.1,5,10),\n",
    "                                            cv=cv)\n",
    "\n",
    "param_range =  np.linspace(0.1,5,10)\n",
    "\n",
    "train_scores_mean =  np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(valid_scores, axis=1)\n",
    "test_scores_std = np.std(valid_scores, axis=1)\n",
    "\n",
    "plt.title(\"Validation Curve SVM - Liver\")\n",
    "plt.xlabel(\"C\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.ylim(0.0, 1.1)\n",
    "lw = 2\n",
    "# semilogx\n",
    "plt.plot(param_range, train_scores_mean, label=\"Training score\",\n",
    "             color=\"darkorange\", lw=lw)\n",
    "\n",
    "plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "plt.plot(param_range, test_scores_mean, label=\"Cross-validation score\",\n",
    "             color=\"navy\", lw=lw)\n",
    "plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                 color=\"navy\", lw=lw)\n",
    "plt.legend(loc=\"best\")\n",
    "\n",
    "\n",
    "plt.savefig(\"SVM_complexity_liver_C.pdf\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "surface-validity",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# cv = ShuffleSplit(n_splits=10, test_size=0.1, random_state=0)\n",
    "cv = RepeatedStratifiedKFold(n_splits=10, n_repeats=2, random_state=45)\n",
    "train_scores, valid_scores = validation_curve(liver_pipeline_rbf,\n",
    "                                              X_train,\n",
    "                                              y_train, \n",
    "                                              param_name =  \"model__gamma\",\n",
    "                                            param_range = np.linspace(1e-2,2.,10),\n",
    "                                            cv=cv)\n",
    "\n",
    "param_range =  np.linspace(1e-2,2.,10)\n",
    "\n",
    "train_scores_mean =  np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(valid_scores, axis=1)\n",
    "test_scores_std = np.std(valid_scores, axis=1)\n",
    "\n",
    "plt.title(\"Validation Curve SVM - Liver\")\n",
    "plt.xlabel(\"Gamma\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.ylim(0.0, 1.1)\n",
    "lw = 2\n",
    "# semilogx\n",
    "plt.semilogx(param_range, train_scores_mean, label=\"Training score\",\n",
    "             color=\"darkorange\", lw=lw)\n",
    "\n",
    "plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "plt.semilogx(param_range, test_scores_mean, label=\"Cross-validation score\",\n",
    "             color=\"navy\", lw=lw)\n",
    "plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                 color=\"navy\", lw=lw)\n",
    "plt.legend(loc=\"best\")\n",
    "plt.savefig(\"SVM_complexity_liver_gamma.pdf\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "pointed-marsh",
   "metadata": {},
   "source": [
    "## Grid Search Liver "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "documented-avatar",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best: 0.885576 usin g {'model__C': 1, 'model__class_weight': 'balanced', 'model__kernel': 'linear', 'model__probability': True}\n"
     ]
    }
   ],
   "source": [
    "# Set the parameters by cross-validation\n",
    "\n",
    "liver_steps =    [('scaler', StandardScaler()),(\"model\",  SVC( cache_size=10000, probability=True))]\n",
    "\n",
    "liver_pipeline= Pipeline(steps=liver_steps)\n",
    "\n",
    "tuned_parameters = [{'model__kernel': ['rbf'], 'model__gamma': np.linspace(1e-2,.1,5),\n",
    "                     'model__C': [ 1,2,3 ], \"model__class_weight\":[\"balanced\", None]},\n",
    "                    {'model__kernel': ['linear'], 'model__C': [ 1,2,3 ], \"model__class_weight\":[\"balanced\", None], \n",
    "                    \"model__probability\":[True]}]\n",
    "\n",
    "\n",
    "# define the evaluation procedure\n",
    "cv = RepeatedStratifiedKFold(n_splits=5, n_repeats=1, random_state=45)\n",
    "# define the grid search procedure\n",
    "grid_search = GridSearchCV(estimator=liver_pipeline, param_grid=tuned_parameters, \n",
    "                           n_jobs=-1, cv=cv, \n",
    "                           scoring='average_precision')\n",
    "# execute the grid search\n",
    "grid_result_l = grid_search.fit(X_train, y_train)\n",
    "# summarize the best score and configuration\n",
    "print(\"Best: %f usin g %s\" % (grid_result_l.best_score_, grid_result_l.best_params_))\n",
    "# summarize all scores that were evaluated\n",
    "means = grid_result_l.cv_results_['mean_test_score']\n",
    "stds = grid_result_l.cv_results_['std_test_score']\n",
    "params = grid_result_l.cv_results_['params']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "unexpected-calibration",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9040576395547891\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_train, liver_pipeline.fit(X_train, y_train).predict_proba(X_train)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "religious-attraction",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8945287189989907\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_train, grid_result_l.predict_proba(X_train)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "multiple-restriction",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9268875935569446\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_test, grid_result_l.predict_proba(X_test)[:,1]))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "located-fourth",
   "metadata": {},
   "source": [
    "## Final learning curve liver svm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "employed-blanket",
   "metadata": {},
   "outputs": [],
   "source": [
    "S"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "southern-genre",
   "metadata": {},
   "outputs": [],
   "source": [
    "liver_steps_f =    [('scaler', StandardScaler()),(\"model\", \n",
    "                                                SVC(C=1, cache_size=10000, class_weight='balanced', kernel='rbf', gamma=0.039 ),) ]\n",
    "\n",
    "liver_pipeline_f= Pipeline(steps=liver_steps_f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "adjustable-future",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "my_learn_curve(title=\" Support Vector Machine\", \n",
    "               estimator=liver_pipeline_f.fit(X_train,y_train), \n",
    "               cv=stratified_kfold, \n",
    "               X=X_train,\n",
    "               y=y_train,\n",
    "               label_suffix=\"Final SVM\", \n",
    "               color_=[\"orange\",\"blue\"],\n",
    "               shape = 'o-'  )\n",
    "\n",
    "# my_learn_curve(title=\" Support Vector Machine - Liver\", \n",
    "#                estimator=liver_pipeline_linear.fit(X_train,y_train), \n",
    "#                cv=stratified_kfold,\n",
    "#                X=X_train,\n",
    "#                y=y_train ,\n",
    "#                label_suffix=\"Base SVM\", \n",
    "#                color_=[\"orange\",\"blue\"], shape = 'x-' )\n",
    "\n",
    "plt.savefig(\"SVM_LIVER_FINAL_learning_curve.pdf\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "substantial-penguin",
   "metadata": {},
   "source": [
    "# Metrics "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "aggressive-broadcasting",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.44      0.92      0.60       138\n",
      "           1       0.94      0.50      0.65       318\n",
      "\n",
      "    accuracy                           0.63       456\n",
      "   macro avg       0.69      0.71      0.63       456\n",
      "weighted avg       0.79      0.63      0.64       456\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.35      0.96      0.52        26\n",
      "           1       0.98      0.48      0.64        88\n",
      "\n",
      "    accuracy                           0.59       114\n",
      "   macro avg       0.66      0.72      0.58       114\n",
      "weighted avg       0.83      0.59      0.61       114\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# print(grid_result_l.best_estimator_.get_params())\n",
    "print(classification_report(y_train, liver_pipeline_f.fit(X_train,y_train).predict(X_train)))\n",
    "print(classification_report(y_test,  liver_pipeline_f.predict(X_test) ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "arabic-clearance",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, ' SVM - Confusion Matrix Liver')"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(estimator=   liver_pipeline_f.fit(X_train,y_train), X=X_train, y_true= y_train  , \n",
    "                      cmap=plt.cm.Blues,\n",
    "                       normalize=\"true\",\n",
    "                      display_labels = [\"Not Patient\" ,\"Patient\"]\n",
    "                        )\n",
    "\n",
    "plt.title(\" SVM - Confusion Matrix Liver\")\n",
    "\n",
    "# plt.savefig(\"SVM_LIVER_BASE_CONFUSION_MAT.PDF\")\n",
    "### Overfit "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "supported-warehouse",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8088284273460264\n",
      "0.8748538011695906\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_train, grid_result_l.predict(X_train)))\n",
    "print(average_precision_score( y_test,grid_result_l.predict(X_test)))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "common-upgrade",
   "metadata": {},
   "source": [
    "## Census Data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "stock-concrete",
   "metadata": {},
   "outputs": [],
   "source": [
    "StandardScaler?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "industrial-grammar",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data shape is (26029, 15)\n",
      "Test data shape is (6508, 15)\n",
      "Gradient Boosting  AUC imputing missing values: 0.893\n"
     ]
    }
   ],
   "source": [
    "census_df = pd.read_csv(f\"{parent_dir}/data/adult.csv\")\n",
    "census_df.drop_duplicates(inplace=True)\n",
    "# # Deleting duplicates \n",
    "census_df.duplicated().sum()\n",
    "#Replacing '?' for nans\n",
    "census_df = census_df.replace('?', np.NaN)\n",
    "# Mapping already our target variable in  1 and 0.\n",
    "census_df.income = census_df.income.map({'<=50K':-1, '>50K':1})\n",
    "\n",
    "train_data, test_data = train_test_split(census_df,\n",
    "                                         test_size=0.2,\n",
    "                                         stratify=census_df['income'],\n",
    "                                         random_state=65)\n",
    "print('Train data shape is {}'.format(train_data.shape))\n",
    "print('Test data shape is {}'.format(test_data.shape))\n",
    "X_c_train = train_data.drop('income', axis=1)\n",
    "y_c_train = train_data['income']\n",
    "\n",
    "#Transforming categorical columns\n",
    "categorical_columns= X_c_train.select_dtypes(object).columns\n",
    "\n",
    "\n",
    "#         ('over', SMOTE()),\n",
    "# Using pipeline\n",
    "census_steps =[('encoding', OneHotEncoder(cols=categorical_columns)),\n",
    "         ('imputer',SimpleImputer(missing_values=np.NaN, strategy='most_frequent')),\n",
    "         ('scaler', StandardScaler()),\n",
    "          ('under', RandomUnderSampler(random_state=62)),\n",
    "        (\"model\",  SVC(gamma='scale', cache_size=10000))]\n",
    "\n",
    "census_pipeline = Pipeline(steps=census_steps)\n",
    "\n",
    "\n",
    "census_steps_rbf = [('encoding', OneHotEncoder(cols=categorical_columns)),\n",
    "         ('imputer',SimpleImputer(missing_values=np.NaN, strategy='most_frequent')),\n",
    "         ('scaler', StandardScaler()),\n",
    "        (\"model\",  SVC(gamma='scale',kernel='rbf', cache_size=10000))]\n",
    "census_pipeline_rbf = Pipeline(steps=census_steps_rbf)\n",
    "\n",
    "census_steps_linear = [('encoding', OneHotEncoder(cols=categorical_columns)),\n",
    "         ('imputer',SimpleImputer(missing_values=np.NaN, strategy='most_frequent')),\n",
    "         ('scaler', StandardScaler()),\n",
    "        (\"model\",  SVC(gamma='scale',kernel='linear', cache_size=10000))]\n",
    "\n",
    "census_pipeline_linear = Pipeline(steps=census_steps_linear)\n",
    "\n",
    "#inital shot \n",
    "stratified_kfold = StratifiedKFold(n_splits = 5, random_state=65, shuffle=True)\n",
    "scores = cross_val_score(census_pipeline, X_c_train, y_c_train, scoring='roc_auc', cv=stratified_kfold, n_jobs=3)\n",
    "print(f'Gradient Boosting  AUC imputing missing values: {round(np.mean(scores),3)}')\n",
    "X_c_test = test_data.drop('income', axis=1)\n",
    "y_c_test = test_data['income']\n",
    "fit_census = census_pipeline.fit(X_c_train,y_c_train)\n",
    "\n",
    "fit_census.score(X_c_test,y_c_test )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "standing-mineral",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "my_learn_curve(title=\" Support Vector Machine\", \n",
    "               estimator=census_pipeline_linear, \n",
    "               cv=stratified_kfold,\n",
    "               X=X_c_train, \n",
    "               y=y_c_train,  \n",
    "               label_suffix=\"linear\", \n",
    "               color_=[\"orange\",\"blue\"],\n",
    "               shape = 'o-'  )\n",
    "\n",
    "plt.savefig(\"SVM_CENSUS_BASE_LEARNING_CURVE.pdf\")\n",
    "\n",
    "\n",
    "# my_learn_curve(title=\" Support Vector Machine\", \n",
    "#                estimator=census_pipeline_rbf, \n",
    "#                cv=stratified_kfold,\n",
    "#                X=X_c_train, \n",
    "#                y=y_c_train,  \n",
    "#                label_suffix=\"rbf\", \n",
    "#                color_=[\"orange\",\"blue\"],\n",
    "#                shape = 'x-'  )\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "id": "controlled-prague",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f6746da4d90>"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWYAAAEGCAYAAABW0j9MAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAAjYUlEQVR4nO3deZhU1Z3/8fenG5BVooKgIIKKC25EEZUoEjQGHOOCG4bEJfFRHNH8nGg0ziLRZEw0jjERh6AxxrgQd9HgkjiaqKhhERdQHAQXQBQQF5BRuvn+/ri3sbrorq6Wru7b1Of1PPfxLqe+91QVfvvUueeeq4jAzMyyo6KlK2BmZrU5MZuZZYwTs5lZxjgxm5lljBOzmVnGtGnpCmxKtFmXqOi0VUtXwxphQJ8tW7oK1kivvPjC8ojovjExKjffPqJqTVFlY82yRyNixMacr7GcmJtQRaetaH/opS1dDWuE+yaMbukqWCP179HxrY2NEVVr2GyXE4sq+3+zJ3Tb2PM1lhOzmZUhgbLbk+vEbGblR0BFZUvXol5OzGZWnqSWrkG9nJjNrAy5K8PMLHvcYjYzyxDhFrOZWbbILWYzs8zxqAwzsyzxxT8zs2wR7sowM8sct5jNzLLEXRlmZtkioNIX/8zMssV9zGZmWeKuDDOz7HGL2cwsY9xiNjPLEPmWbDOz7MnwLdnZbcubmZVMevGvmKWhSNIISfMkzZd0cR3Hu0p6UNKLkuZIOr2hmE7MZlaearozGloKhlAlMAEYCQwATpY0IK/YOcDciNgbGAZcLaldobhOzGZWfmrmY974FvNgYH5ELIiIz4HJwNF5ZQLoIklAZ+ADoKpQUPcxm1kZatQ45m6SZuRsT4qISel6L+CdnGOLgP3zXn8dMAVYAnQBToqIdYVO6MRsZuWp+It/yyNiUD3H6urriLztbwKzgeHAjsBfJD0VER/XW7Via2Zmtklpgj5mkhbydjnbvUlaxrlOB+6NxHxgIbBroaBOzGZWftRkozKmA/0l9Usv6I0m6bbI9TZwaHJa9QB2ARYUCuquDDMrT01wg0lEVEkaBzwKVAI3RcQcSWPT4xOBy4GbJb1M0vVxUUQsLxTXidnMypKa6M6/iJgKTM3bNzFnfQlweGNiOjGbWdlJnizlW7LNzLJDQhVOzGZmmeIWs5lZxjgxm5lljBOzmVmWiLrv2csIJ2YzKztCbjGbmWVNRUV2b3x2YjazsuQWs5lZlriP2cwse9xiNjPLEF/8MzPLIN+SbWaWJXJXhplZ5jgxm5lljBOzmVmG+OKfmVkWZTcvOzGbWRlStm/Jzm7NzMxKSFJRSxFxRkiaJ2m+pIvrOH6hpNnp8oqkaklbForpxGxm5UlFLoVCSJXABGAkMAA4WdKA3DIRcVVEDIyIgcCPgb9FxAeF4roro8wN37sXV5yyPxUV4tYnXufaKS/XOj7uyD04/ms7ANCmsoKde3Vl5zPv4MPVnzN25AC+O3xnImDuOys5d+LTfLa2uiXeRll5evpr/GLiFKqr1zFq5GDOOGl4reMP/c8sbrrzCQA6tt+Mfz93FLvsuC1L3/+QS66azPKVn1AhcfwR+/OdYw9uibeQCU108W8wMD8iFqQxJwNHA3PrKX8ycEdDQUvWYpYUkq7O2b5A0vgGXnNM/l+bnGPjJS3O+TlwVAOxLsnbntaI6ufHOk3Stl/29VlVIXHl6Qdw4i8eY8gF9zFqyA7s0qtrrTLXPfQKw348hWE/nsLlk2cy7dX3+HD152yzRUfOHDGAQy95kIN+dD+VFWLUgf1a6J2Uj+rqdfxswn1c/9Pv88ANF/DwE7N54633apXp3WNLfn/V2dw78YecNeYwfnLt3QBUVlZwwZlHMuXGC7nt2nFMfnDaBq8tF8V2YxSRvHsB7+RsL0r31XXOjsAI4J6GgpayK+MzYJSkbo14zTEkPwfqc036c+AE4CZJhepfKzFHxJBG1CPfacAml5j32akbC5d+wlvvr2Jt9True3YBIwf1qbf8qCH9uGfagvXbbSoraN+uksoK0aFdG95d+WlzVLusvTzvbfps243tttmKtm3bMHLYQJ54dk6tMgN370vXLh0B2GvXPry3/CMAum+1OQP69wagU8f29Ntu6/XHylEjEnM3STNyljNzw9QROuo55beAZxrqxoDSJuYqYBJwfv4BSdtLelzSS+l/+0gaAhwFXJW2inesL3BEvJrG7ybpfkkzJc2p+cAk/RzokMa5Ld23Kuf8F0qanp7/J+m+vpJelXRDGusxSR0kHQ8MAm5L43Vosk+ohW2zRUcWr1i9fnvJik/ZZotOdZbt0K6SQ/fuzYPPvwnAuys/5bqHXuHF605k7n+P5uNPP+fJl5c0R7XL2vsrPqZn96+s3+7RrWvB5HrfI//goP123WD/4qUf8NobS9hr1/r/EG/qVKGiFmB5RAzKWSblhFkEbJez3Ruo73+E0RTRjQGlv/g3ARgjqWve/uuAWyJiL+A24NcRMQ2YAlyYdpS/UV9QSfsD64BlwPciYl+S5HmepK0i4mJgTRpnTN5rDwf6k/QNDQT2lTQ0PdwfmBARuwMfAsdFxN3ADGBMGm9NXrwza/6SxmeraE3q+pkW9fyx/+Y+fXh+XtKNAdC1UzuOGNSHfc67i93/eTKdNmvDCQftUNL6GkRs+P3U93P7H7Pnc++j0zn/+0fU2v/pms84//JbuGjsUXTu1L4k9WwNmqgrYzrQX1I/Se1Iku+UOs7VFTgEeKCYupU0MUfEx8AtwHl5hw4Ebk/X/wgcVGTI8yXNBn4JnBTJv9LzJL0IPEfyl6t/AzEOT5cXgFnArjmvWRgRs9P1mUDfhioUEZNq/pJqs85Fvo1sWPLBanpt9UULedutOrK0nu6IUUP6ce+0heu3D9ljW956/xNWfPIZVdXBQ9PfYvDOW5e8zuWuR7euLF324frt95Z/xNZbbb5BuXkLlnDpr+7i1+NP4yubf/Edr62q5vzLb+Gfhn+Vww7aszmqnE1qmsQcEVXAOOBR4FXgzoiYI2mspLE5RY8FHouI1XXFydccw+V+BXwfqPs3cqK+Ppl816St1oMj4ilJw4DDgAMjYm+SZNtQE0DAFTXDVyJip4j4XXrss5xy1Wzio1ZeeGM5O/TcnD7dO9O2soJjD9yBh2e+s0G5Lh3aMmS3njw88+31+xYvX8Wg/t3p0K4SgKF7bMvri8u3v7K57LHLdry1eDmLln7A2rVVPPzkbIYdUPuyzLvvr+T8y27higtPpm/v7uv3RwSX/ted7LDd1px63CHNXfVMESAVtzQkIqZGxM4RsWNE/CzdNzEiJuaUuTkiRhdbv5Innoj4QNKdJMn5pnT3NJIm/x+BMcDT6f5PgC6NCN8VWBkRn0raFTgg59haSW0jYm3eax4FLpd0W0SsktQLyC+Tr7H1ahWq1wUX3fwcd/34cCorxO1P/i/zFn3IaYftAsDNf50HwJH7bc8TLy3m08+q1r925hvLmfL8mzzxn0dRtS54+c0V/OHxeS3yPspJm8pKLjnnGMZecgPV69Zx7OGD2alvT+586FkATjzyQCbe9lc+/ORTfnrdvQBUVlbyp+t+wAtz3uTBx2fRv19Pjj/7vwA47/SRDB28W4u9n5aT7bkyVFefVZMEllZFROd0vQewELgyIsZL6kuSpLuR9BOfHhFvS/oacANJy/X43H7mdKjdqoj4Zc6+zYD7SYanzAO6A+Mj4klJvyC5mDgrIsbk1ecHwBlpmFXAd0hayA9FxB5pmQuAzml9jwP+E1hD0jqv1c9co3LLvtH+0Es36nOz5vXihKIbMZYR/Xt0nBkRgzYmRvueO8f2p/6mqLKvXzlio8/XWCVrMdckwXT9PaBjzvabwPA6XvMM9QyXi4jxdez7jOSOm7rKXwRcVE99rgWureNle+SU+WXO+j0UMfbQzFqJIrspWsom3YdqZlYXARV+tJSZWba4xWxmljFZvvjnxGxm5cd9zGZm2SKU6YnynZjNrCy5xWxmljHuYzYzyxL3MZuZZUsyV0Z2M7MTs5mVpQznZSdmMytPvvPPzCxL5K4MM7NMqZmPOaucmM2sDGV7PmYnZjMrSxnOy07MZlaG5It/ZmaZkvVxzNmdxcPMrISa4inZaZwRkuZJmi/p4nrKDJM0W9IcSX9rKKZbzGZWlpqiwSypEpgAfANYBEyXNCUi5uaU+QpwPTAifbbp1g3FdYvZzMpSE7WYBwPzI2JBRHwOTAaOzivzbeDeiHgbICLebyioE7OZlZ90EqNiFqCbpBk5y5k5kXoB7+RsL0r35doZ2ELSk5JmSjqloeq5K8PMyk4yUX7RfRnLI2JQvaE2FHnbbYB9gUOBDsCzkp6LiNfrO6ETs5mVpYqmGZWxCNguZ7s3sKSOMssjYjWwWtLfgb2BehOzuzLMrCw1oiujkOlAf0n9JLUDRgNT8so8ABwsqY2kjsD+wKuFgrrFbGZlR000iVFEVEkaBzwKVAI3RcQcSWPT4xMj4lVJjwAvAeuAGyPilUJxnZjNrCw11Y1/ETEVmJq3b2Le9lXAVcXGrDcxS/oNG3Zi557ovGJPYmaWNa31luwZzVYLM7NmJJKRGVlVb2KOiD/kbkvqlF5VNDNr9TLcYG54VIakAyXNJb2KKGlvSdeXvGZmZqVS5F1/LTXRUTHD5X4FfBNYARARLwJDS1gnM7OSa6LhciVR1KiMiHgn7y9HdWmqY2ZWeqLJbjApiWIS8zuShgCRDqA+jwYGR5uZZV2WR2UU05UxFjiHZGKOxcDAdNvMrFUqthsjs10ZEbEcGNMMdTEzazZZ7sooZlTGDpIelLRM0vuSHpC0Q3NUzsysVFTk0hKK6cq4HbgT2AbYFrgLuKOUlTIzK7XWPlxOEfHHiKhKl1spcKu2mVnWJaMyiltaQqG5MrZMV59IHzA4mSQhnwT8uRnqZmZWGmrURPnNrtDFv5kkibim9mflHAvg8lJVysys1Fqqm6IYhebK6NecFTEzay41XRlZVdSdf5L2AAYA7Wv2RcQtpaqUmVmptcoWcw1JlwLDSBLzVGAk8DTgxGxmrVZ203JxozKOJ3m669KIOJ3kIYKblbRWZmYlJEFlhYpaWkIxXRlrImKdpCpJmwPvA77BxMxatSx3ZRTTYp4h6SvADSQjNWYB/yhlpczMSq2p5sqQNELSPEnz06HF+ceHSfpI0ux0+Y+GYhYzV8Y/p6sT0ye9bh4RLzVcXTOzbBJqkrkyJFUCE4BvAIuA6ZKmRMTcvKJPRcSRxcYtdIPJPoWORcSsYk9iZpYpTTdz3GBgfkQsAJA0GTgayE/MjVKoxXx1gWMBDN+YE2+KBvbrxjN3nN7S1bBG2GK/cS1dBWshjehj7iYp9+HUkyJiUrreC3gn59giYP86Yhwo6UVgCXBBRMwpdMJCN5h8vbg6m5m1LgIqi0/MyyNiUIFQ+fLnEpoFbB8RqyQdAdwP9C90wmIu/pmZbXKaaBKjRcB2Odu9SVrF60XExxGxKl2fCrSV1K1g3Rr9bszMNgFNlJinA/0l9UsfvTcamJJbQFJPpf0mkgaT5N0VhYIWdUu2mdmmJBkKt/FX/yKiStI44FGgErgpIuZIGpsen0hyk97ZkqqANcDoiCg4dXIxt2SL5NFSO0TEZZL6AD0jwmOZzazVaqqb+tLuial5+ybmrF8HXNeouhVR5nrgQODkdPsTknF7ZmatVqt+GCuwf0TsI+kFgIhYmfalmJm1SgLaZPiW7GIS89r07pYAkNQdWFfSWpmZlViG83JRifnXwH3A1pJ+RtKR/W8lrZWZWQlJTXNLdqkUM1fGbZJmkkz9KeCYiHi15DUzMyuhDOflokZl9AE+BR7M3RcRb5eyYmZmpdTaHy31Z754KGt7oB8wD9i9hPUyMysZQYtNgl+MYroy9szdTmedO6ue4mZm2VfcXX0tptF3/kXELEn7laIyZmbNRRl+6l8xfcz/krNZAewDLCtZjczMSky0/hZzl5z1KpI+53tKUx0zs+bRahNzemNJ54i4sJnqY2bWLLL8MNZCj5Zqk86cVO8jpszMWiMJKjM86XGhFvM/SPqTZ0uaAtwFrK45GBH3lrhuZmYl06rv/AO2JJnUeThfjGcOwInZzFql1nzxb+t0RMYrfJGQaxSc5NnMLOsy3GAumJgrgc4U97BBM7NWRFS00nHM70bEZc1WEzOzZiJab4s5w9U2M9sIgjYZ7mQuNGDk0GarhZlZM6ppMTfFo6UkjZA0T9J8SRcXKLefpGpJxzcUs94Wc0R80HCVzMxap6YYLpfehDcB+AawCJguaUpEzK2j3C9InqbdcN02umZmZq1QE7WYBwPzI2JBRHwOTAaOrqPcuSRTWbxfTN2cmM2s7Igk+RWzAN0kzchZzswJ1Qt4J2d7Ubrvi3NJvYBjgYnF1q/R036ambV6alRXxvKIGFR/pA3kDyf+FXBRRFQXOz+HE7OZlZ3kzr8mGZWxCNguZ7s3sCSvzCBgcpqUuwFHSKqKiPvrC+rEbGZlqYkGy00H+kvqBywGRgPfzi0QEf3Wn1O6GXioUFIGJ2YzK1NN0WBOZ+AcRzLaohK4KSLmSBqbHi+6XzmXE7OZlSE12XzMETEVmJq3r86EHBGnFRPTidnMyk7NqIyscmI2s7LU2udjNjPbtKiVPlrKzGxT5a4MM7MMcovZzCxjspuWnZjNrAwJqHSL2cwsWzKcl52YzawcCWW4M8OJ2czKklvMZmYZkgyXy25mdmI2s/JT5PP8WooTs5mVJd+SbWaWIclE+S1di/o5MZtZWfKoDDOzjMlwT4YTc7n767S5/Pjqu6let47vHj2E8087vNbx199cyrjLbuXF1xbxb2cfybnfPQyARUtXcvb4W3h/xcdUSJx67NcYe/LXW+ItlJ1DD9yNK354PJUVFfzxgWn86g9/qXV8807t+e3lp9K7xxZUtqnkulsf5/YHnwPg7JO/znePGQIRzJ2/hHMuu5XPPq9qibfR4rLcYs7yBEsbkFQtabakVyTdJaljgbLDJA3J2R4r6ZQved6+kr7dcMnWpbp6HRdeeSd3XfvPPHfnv3HPYzN5bcG7tcpssXknfv7DExj3neG19rdpU8FP/98onr/r33ns9xdw491/3+C11vQqKsRVPzqRE35wPQec+FOOO3xfdunXs1aZM04YyrwFSzl4zM/51lnX8tMfHEvbNpVs070rZ510CMNPuZIho/+TiooKRh2+bwu9k5ZV08dczNISWlViBtZExMCI2AP4HBhboOwwYH1ijoiJEXHLlzxvX/IesLgpmDnnTXbYrht9e3ejXds2jPrGPkz920u1ynTfsgv77L49bdtU1trfs1tX9t41eThwl07t2blvT95d9mFzVb1s7bt7Xxa8s5y3Fq9gbVU19/5lFkccsletMgF07rQZAJ06bsbKjz+lqnodAG3aVNJ+s7ZUVlbQsX07li77qLnfQjZIVBS5tITWlphzPQXsJOlbkp6X9IKkv0rqIakvSdI+P21hHyxpvKQLACTtKOkRSTMlPSVp13T/zZJ+LWmapAWSjk/P9XPg4DTW+S3xZkvh3WUf0avHFuu3t+2xBe9+if9R316ygpfmLWLf3fs2Ye2sLtt078ri91au317y3kq26d61Vpkb7vwbO/ftyasP/4xn7riEH199NxHBu8s+4je3Ps7LD17Oaw//jI9Xr+GJ519r7reQGSpyaTCONELSPEnzJV1cx/GjJb2U5o8Zkg5qKGarTMyS2gAjgZeBp4EDIuKrwGTgRxHxJjARuCZtYT+VF2IScG5E7AtcAFyfc2wb4CDgSJKEDHAx8FQa65q8upyZftgzli1f1qTvs9QiYoN9jW0grPr0M0656Eau+Jfj2LxzhyaqmdWnrjmE87/G4QfsxsuvL2K3kf/K0DFXcOWFJ9ClU3u6dunAEUP3ZODRl7LbyH+lY/t2nDhyv2aqebYkXRkb32KWVAlMIMlHA4CTJQ3IK/Y4sHdEDAS+B9zYUP1aW2LuIGk2MAN4G/gd0Bt4VNLLwIXA7oUCSOpM0sVxVxrrtyTJuMb9EbEuIuYCPRqqUERMiohBETGoe7fuX+IttZxtt/7KBq2vnt26FnhFbWurqjn1ohs4YcQgvjV8YAlqaPmWvP/hBr9yli6v/StnzLcO4KEnXgRg4aLlvLVkBf2378Gwwbvy1pIVrPhwFVXV63jwiRcZvFe/Zq1/ljRRi3kwMD8iFkTE5ySNw6NzC0TEqviiFdSJpLepoNaWmGv6mAdGxLnpB/Eb4LqI2BM4C2jfQIwK4MOcOAMjYrec45/lrGf3sm0T2GfA9rzx9jLeWrycz9dWce9fZjFy6F4Nv5CktX3u5bexc9+enDPm0BLX1GrMmvsWO/bpTp9tt6Jtm0pGfWMfHv577esCi5auZOh+uwDJNYKdtu/Bm4uXs2jpBwzasx8dNmsLwCH77cK8he81+3vIjKbJzL2Ad3K2F6X7ap9KOlbSa8CfSVrNBW0Kw+W6AovT9VNz9n8CbJ5fOCI+lrRQ0gkRcZeS34Z7RcSLBc7xCdClyWqcEW3aVHLlj07kuPMmUF0djDnqAHbbcRtuuifp+fnecQfz3vKPGX7qlXyy+v+QxMTJT/Lsn/6VOfOX8Kep/2DATtty8LevAODfzzmKw79W8AeLbaTq6nX86Mo7uefX51BZKW6b8hyvLVjK6aOSbsvf3/s0V/3uESZc+h2eueMSJPjJdQ/wwUer+eCj1Ux5/AWevPUiqqvX8dK8Rfzhvmda+B21nEZc2OsmaUbO9qSImJSu1xVkgxZxRNwH3CdpKHA5cFihE6qufsaskrQqIjrn7TsauIYkOT8H7BcRwyTtDNwNrAPOBQ4FVkXELyX1A/6bpAujLTA5Ii6TdDPwUETcnXs+SW2BR4BuwM35/cw19t13UDzz/Iy6DllGbbHfuJaugjXS/82eMDMiBm1MjN32/Grc8sCTRZUdvONX6j2fpAOB8RHxzXT7xwARcUV98SQtJMlTy+sr06pazPlJOd33APBAHftfB3J/lz+Vc2whMKKO15xW1/kiYi1JYjezTUXTdFROB/qnjb3FwGjyhtZK2gl4IyJC0j5AO2BFoaCtKjGbmTWFpPt44zNzRFRJGgc8ClQCN0XEHElj0+MTgeOAUyStBdYAJ0UDXRVOzGZWfppwPuaImApMzds3MWf9F8AvGhPTidnMylKWh1w5MZtZGVKdN+tkhROzmZWlDOdlJ2YzKz/FzoPRUpyYzaw8ZTgzOzGbWVnK8kT5TsxmVpbcx2xmliVNOI65FJyYzawsuSvDzCxDhFvMZmaZk+G87MRsZmUqw5nZidnMylJLPQG7GE7MZlaWspuWnZjNrFxlODM7MZtZ2WmqifJLxYnZzMqPbzAxM8ueDOdlJ2YzK0eeKN/MLHMynJepaOkKmJk1NzViaTCWNELSPEnzJV1cx/Exkl5Kl2mS9m4oplvMZlaemqDFLKkSmAB8A1gETJc0JSLm5hRbCBwSESsljQQmAfsXiuvEbGZlqYmGyw0G5kfEAgBJk4GjgfWJOSKm5ZR/DujdUFB3ZZhZWZKKW4BukmbkLGfmhOkFvJOzvSjdV5/vAw83VDe3mM2s/Agqim8wL4+IQfVH2kDUWVD6OkliPqihEzoxm1mZapKujEXAdjnbvYElG5xJ2gu4ERgZESsaCuquDDMrOzUT5RfZlVHIdKC/pH6S2gGjgSm1ziX1Ae4FvhsRrxdTP7eYzawsNUV7OSKqJI0DHgUqgZsiYo6ksenxicB/AFsB16c3tVQV6BoBnJjNrEw11Q0mETEVmJq3b2LO+hnAGY2J6cRsZmXJt2SbmWVMdtOyE7OZlaEiL+y1GCdmMytLnijfzCxrspuXnZjNrDxlOC87MZtZORIVGe5kdmI2s7JTc+dfVvmWbDOzjHGL2czKUpZbzE7MZlaWPFzOzCxLfIOJmVm2ZP3inxOzmZUld2WYmWWMW8xmZhmT4bzsxGxmZSrDmdmJ2czKjiDTt2Qros4nbduXIGkZ8FZL16NEugHLW7oS1iib6ne2fUR035gAkh4h+XyKsTwiRmzM+RrLidmKImlGQw+QtGzxd9Z6ea4MM7OMcWI2M8sYJ2Yr1qSWroA1mr+zVsp9zGZmGeMWs5lZxjgxm5lljBNzKyYpJF2ds32BpPENvOYYSQPqOTZe0mJJsyW9IumoBmJdkrc9rRHVz491mqRtv+zrNzWSqnO+h7skdSxQdpikITnbYyWd8iXP21fSt7/Ma63pODG3bp8BoyQVO1Ae4BigzsScuiYiBgInADdJKvRvpFZijogh9RUswmmAE/MX1kTEwIjYA/gcGFug7DBg/WcfERMj4pYved6+gBNzC3Nibt2qSK68n59/QNL2kh6X9FL63z5pq+oo4Kq0NbZjfYEj4tU0fjdJ90uaKWmOpDPT+D8HOqRxbkv3rco5/4WSpqfn/0m6r6+kVyXdkMZ6TFIHSccDg4Db0ngdmuwT2jQ8Bewk6VuSnpf0gqS/SuohqS9J0j4//ewOTn/5XAAgaUdJj6Tf31OSdk333yzp15KmSVqQfgcAPwcOTmNt8O/KmklEeGmlC7AK2Bx4E+gKXACMT489CJyarn8PuD9dvxk4vp5444EL0vX9gSUk0wpsme7rALwCbFVz/vz6pP89nOQPhkj++D8EDCVpjVUBA9NydwLfSdefBAa19GealSXns2wDPACcDWzBFyOpzgCuzv/e6vgeHwf653yn/5Pz7+Cu9PsZAMxP9w8DHmrp91/uiycxauUi4mNJtwDnAWtyDh0IjErX/whcWWTI8yV9B/gEOCkiQtJ5ko5Nj28H9AdWFIhxeLq8kG53Tl/zNrAwIman+2eSJGvbUAdJs9P1p4DfAbsAf5K0DdAOWFgogKTOJF0cd+mLCXs2yylyf0SsA+ZK6tGEdbeN5MS8afgVMAv4fYEyxQ5YvyYiflmzIWkYcBhwYER8KulJoH0DMQRcERG/rbUz+dn9Wc6uapJWuG1oTSR9/etJ+g3wXxExJf1exjcQowL4MD9OjtzvIrtTrZUh9zFvAiLiA5Juge/n7J4GjE7XxwBPp+ufAF0aEb4rsDJNyrsCB+QcWyupbR2veRT4XtpiQ1IvSVs3cJ7G1qscdQUWp+un5uyv87OLiI+BhZJOAFBi7wbO4e8hA5yYNx1XU3saw/OA0yW9BHwX+EG6fzJwYXoBqd6LfzkeAdqkcS4Hnss5Ngl4qebiX42IeAy4HXhW0svA3TT8P/vNwERf/CtoPEm3xFPUns7zQeDYmot/ea8ZA3xf0ovAHODoBs7xElAl6UVf/Gs5viXbzCxj3GI2M8sYJ2Yzs4xxYjYzyxgnZjOzjHFiNjPLGCdma1aNmTWtiFg318zxIOnG+mbNS4/XmoGtEed4s65Jourbn1dmVaHjdZRfP8eFlTcnZmtuBWdNk1T5ZYJGxBkRMbdAkWHkzMBmlmVOzNaSamZNGybpCUm3Ay9LqpR0Vc7sdGfB+jvXrpM0V9KfgfV3E0p6UtKgdH2EpFnpTRKP1zMDW3dJ96TnmC7pa+lrt0pnvXtB0m8p4lZl1TH7Xs6xq9O6PC6pe7qvzhnfzGp4rgxrEZLaACNJ7iwEGAzsEREL0+T2UUTsJ2kz4BlJjwFfJZnIZ0+gBzAXuCkvbnfgBmBoGmvLiPhA0kSSGdt+mZa7nWRekKcl9SG5jXw34FLg6Yi4TNI/AbUSbT2+l56jAzBd0j0RsQLoBMyKiB9K+o809jiSOybHRsT/StofuB4Y/iU+RttEOTFbc6tr1rQhwD8ioma2tMOBvXLmCO5KMjvdUOCOiKgGlkj6nzriHwD8vSZWOo9IXQ4DBuTMura5pC7pOUalr/2zpJVFvKf6Zt9bB/wp3X8rcG8RM76ZOTFbs6tr1jSA1bm7gHMj4tG8ckfQ8Cx5KqIMJN14B0ZE7lSpNXUpep6CRs6+FzQ845uZ+5gtkx4Fzq6ZuU7SzpI6AX8HRqd90NsAX6/jtc8Ch0jql752y3R//qxpj5F0K5CWG5iu/p1k4h8kjSSZnL6QQrPvVQA1rf5vk3SRfJkZ36zMODFbFt1I0n88S9IrwG9Jft3dB/wv8DLw38Df8l8YEctI+oXvTWdUq+lKyJ+B7TxgUHpxcS5fjA75CTBU0iySLpW3G6hrodn3VgO7S5pJ0od8Wbq/sTO+WZnx7HJmZhnjFrOZWcY4MZuZZYwTs5lZxjgxm5lljBOzmVnGODGbmWWME7OZWcb8fxchxXVB1kjzAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(estimator= census_pipeline_rbf.fit(X_c_train,y_c_train),\n",
    "                      X=X_c_train,\n",
    "                      y_true= y_c_train  , \n",
    "                      cmap=plt.cm.Blues,\n",
    "                       normalize=\"true\",\n",
    "                      display_labels = [\"Not Patient\" ,\"Patient\"]\n",
    "                        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "numerous-development",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7f672f433ee0>"
      ]
     },
     "execution_count": 136,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(estimator= census_pipeline_linear.fit(X_c_train,y_c_train),\n",
    "                      X=X_c_train,\n",
    "                      y_true= y_c_train  , \n",
    "                      cmap=plt.cm.Blues,\n",
    "                       normalize=\"true\",\n",
    "                      display_labels = [\"Not Patient\" ,\"Patient\"]\n",
    "                        )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "permanent-luxury",
   "metadata": {},
   "source": [
    "## Complexity \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "entitled-lying",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# cv = ShuffleSplit(n_splits=10, test_size=0.1, random_state=0)\n",
    "cv = RepeatedStratifiedKFold(n_splits=5, n_repeats=1, random_state=45)\n",
    "train_scores, valid_scores = validation_curve(census_pipeline_rbf,\n",
    "                                              X_c_train,\n",
    "                                              y_c_train, \n",
    "                                              param_name =  \"model__C\",\n",
    "                                            param_range = np.linspace(0.1,5,10),\n",
    "                                            cv=cv)\n",
    "\n",
    "param_range =  np.linspace(0.1,5,10)\n",
    "\n",
    "train_scores_mean =  np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(valid_scores, axis=1)\n",
    "test_scores_std = np.std(valid_scores, axis=1)\n",
    "\n",
    "plt.title(\"Complexity Curve SVM\")\n",
    "plt.xlabel(\"C Values\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.ylim(0.0, 1.1)\n",
    "lw = 2\n",
    "# semilogx\n",
    "plt.plot(param_range, train_scores_mean, label=\"Training score\",\n",
    "             color=\"darkorange\", lw=lw)\n",
    "\n",
    "plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "plt.plot(param_range, test_scores_mean, label=\"Cross-validation score\",\n",
    "             color=\"navy\", lw=lw)\n",
    "plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                 color=\"navy\", lw=lw)\n",
    "plt.legend(loc=\"best\")\n",
    "\n",
    "\n",
    "plt.savefig(\"complexity_svc_C.png\")\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "consecutive-quarterly",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "noticed-slave",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# cv = ShuffleSplit(n_splits=10, test_size=0.1, random_state=0)\n",
    "cv = RepeatedStratifiedKFold(n_splits=5, n_repeats=1, random_state=45)\n",
    "train_scores, valid_scores = validation_curve(census_pipeline_rbf,\n",
    "                                              X_c_train,\n",
    "                                              y_c_train, \n",
    "                                              param_name =  \"model__gamma\",\n",
    "                                            param_range = np.linspace(1e-3,1.,5),\n",
    "                                            cv=cv)\n",
    "\n",
    "param_range =  np.linspace(1e-3,1.,5)\n",
    "\n",
    "train_scores_mean =  np.mean(train_scores, axis=1)\n",
    "train_scores_std = np.std(train_scores, axis=1)\n",
    "test_scores_mean = np.mean(valid_scores, axis=1)\n",
    "test_scores_std = np.std(valid_scores, axis=1)\n",
    "\n",
    "\n",
    "plt.title(\"Complexity Curve SVM \")\n",
    "plt.xlabel(\"Gamma Values\")\n",
    "plt.ylabel(\"Score\")\n",
    "plt.ylim(0.0, 1.1)\n",
    "lw = 2\n",
    "# semilogx\n",
    "plt.semilogx(param_range, train_scores_mean, label=\"Training score\",\n",
    "             color=\"darkorange\", lw=lw)\n",
    "\n",
    "plt.fill_between(param_range, train_scores_mean - train_scores_std,\n",
    "                 train_scores_mean + train_scores_std, alpha=0.2,\n",
    "                 color=\"darkorange\", lw=lw)\n",
    "plt.semilogx(param_range, test_scores_mean, label=\"Cross-validation score\",\n",
    "             color=\"navy\", lw=lw)\n",
    "plt.fill_between(param_range, test_scores_mean - test_scores_std,\n",
    "                 test_scores_mean + test_scores_std, alpha=0.2,\n",
    "                 color=\"navy\", lw=lw)\n",
    "plt.legend(loc=\"best\")\n",
    "\n",
    "\n",
    "plt.savefig(\"complexity_svc_gamma.pdf\")\n",
    "\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "functional-passenger",
   "metadata": {},
   "source": [
    "## Grid search CV "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "metropolitan-immunology",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set the parameters by cross-validation\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "tuned_parameters = [{'model__kernel': ['rbf'], 'model__gamma': np.linspace(1e-2,.1,5),\n",
    "                     'model__C': [ 1,2,3 ], \"model__class_weight\":[\"balanced\", None]},\n",
    "                    {'model__kernel': ['model__linear'], 'model__C': [ 1,2,3 ]}]\n",
    "\n",
    "\n",
    "# define the evaluation procedure\n",
    "cv = RepeatedStratifiedKFold(n_splits=5, n_repeats=1, random_state=45)\n",
    "# define the grid search procedure\n",
    "grid_search = GridSearchCV(estimator=census_pipeline, param_grid=tuned_parameters, \n",
    "                           n_jobs=-1, cv=cv, \n",
    "                           scoring='average_precision')\n",
    "# execute the grid search\n",
    "grid_result = grid_search.fit(X_c_train, y_c_train)\n",
    "# summarize the best score and configuration\n",
    "print(\"Best: %f usin g %s\" % (grid_result.best_score_, grid_result.best_params_))\n",
    "# summarize all scores that were evaluated\n",
    "means = grid_result.cv_results_['mean_test_score']\n",
    "stds = grid_result.cv_results_['std_test_score']\n",
    "params = grid_result.cv_results_['params']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "periodic-advancement",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Best: 0.735783 usin g \n",
    "#     {'model__C': 1, 'model__class_weight': 'balanced', 'model__gamma': \"scale\", 'model__kernel': 'rbf'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "announced-thanksgiving",
   "metadata": {},
   "outputs": [],
   "source": [
    "params = grid_result.cv_results_['params']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "regulation-bedroom",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "protected-sperm",
   "metadata": {},
   "outputs": [],
   "source": [
    "#         ('over', SMOTE()),\n",
    "# Using pipeline\n",
    "census_steps =[('encoding', OneHotEncoder(cols=categorical_columns)),\n",
    "         ('imputer',SimpleImputer(missing_values=np.NaN, strategy='most_frequent')),\n",
    "         ('scaler', StandardScaler()),\n",
    "        (\"model\",  SVC( cache_size=10000, class_weight=\"balanced\", gamma=\"scale\", kernel=\"rbf\", probability=True))]\n",
    "\n",
    "census_pipeline = Pipeline(steps=census_steps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "accredited-large",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAA2eUlEQVR4nO3deXwURfr48c8zOUEgImjkUsCLMwQMh5wJsqLuivoVFBQV8MIL0cUDdv3Kuvv1QF2PVWEVEQ9WBVn8ee16JgZEJSCXiCBCgIhc4T5yzdTvj+6ZdCYzk4NMAvTzftGv6a6u7q4pJvV0dffUiDEGpZRS7uWp6wIopZSqWxoIlFLK5TQQKKWUy2kgUEopl9NAoJRSLqeBQCmlXE4DgVIqQERGiciCCOv/IyLX12aZVPRpIFDVJiKXisgyEdknIjtF5AsRaV3X5QpWicbtnyLyeoj0FBEpFJGTqnncySLyZnW2jbDPmSJiRGRIUPozdvqomjxeMGPMRcaY16J5DFX7NBCoahGRM4HXgT8CSUAb4EXAV5flCiYisZXINhP4HxE5ISj9OuBDY8yuGi9YJUQo+1rg+qB8w4BfaqNc6vijgUBVVyqwwRjzhbHsN8bMNcZsgsCZ69/8mUUkXUTyHMu5IjJRRH4Ukd0i8qqIJDrzisgku6eRKyLXOLZNEpHXRWSHiGwUkT+LiMdeN0pEvhaRp0VkF/AOMA04T0QOiMie4DdijPkG+BW4wnGMGOBq4DV7eYyIrLbL+omInO7I21FEPhORXSKyzS73hcAk4Cr7uMvtvM1F5H077zoRucmxn8ki8q6IvCki+4BRYer+A6CPiDS2ly8EVgBbHfs6Q0S+FJF8uw5niciJjvWtROTfdh3mi8jzzgOIyJP2e90gIhc50rNE5EZHXS+IkDdJRF4Rkd9E5FcR+Ztdr+ooo4FAVdf3QDu7wc0QkQbV2Mc1wGDgDOBs4M+OdacCTYEWWGe/L4nIOfa6f2D1QtoCA7DO3Ec7tu0JrAdOAUYCY4FvjDENjDEnhinL6/Z+/AYBccB/ROQyrEb9f4CTgfnAWwAi0hD4HPgv0Bw4E/jCGPNf4BHgHfu4Xez9vgXk2XmHAo+IyPmO414KvAucCMwKU9YC4H1guL18nV1+JwEetY/THmgFTLbLHAN8CGwEWmPV8duObXsCa7DqfwrwiohImLJEyvsaUGLXSVfgAuDGMPtRdckYo5NO1ZqAXsBsYAdW4zQTaGCvmwn8zZE3HchzLOcCYx3LFwO/OPKWACc41s8GHgRigEKgg2PdLUCWPT8K2BRUzlHAggrey2lAMdDSXp4FPGvP/we4wZHXAxwCTgdGAEvD7HMy8KZjuRXgBRo60h4FZjryZ1dQzpnA34C+wDdYAXEbUA9YAIwKs91l/nIC59n/Z7Eh8o0C1jmW6wMGONVezgJurCgvkGz/P9VzrB8BZNb151an8pP2CFS1GWO+NcZcaYw5GegH9Af+VIVdbHbMb8Q6e/XbbYw5GGJ9UyDeXnauaxFmv5VirEta2cBIu3dzGfZlIawG/1kR2WNfWtqFdcbdAqtxr+y1+ebALmPM/iMtuzFmAVbv5M9Y9zEOO9eLyCki8rZ9SWYf8CZW3WGXeaMxpiTM7gOXmIwxh+zZcD2+cHlPx+pR/eaot39i9dLUUUYDgaoRxpgc4N9AJzvpINYZot+pITZr5Zg/DdjiWG4cdPPWv34n1pn76UHrfnUWJ7h4FZXf9hrWZZYrsO5/fG+nbwZuMcac6JjqGWMW2uvOCLO/4ONuAU6yLydVtuyRvIl1s77cE09YPQ0DpBhjGmFdIvNfstkMnFbJG+nVtRmrR9DUUWeNjDEdo3hMVU0aCFS1iEhfEblJRE6xl9sBQ4Bv7SzLgItF5CQRORUYH2I3t4tIS7Eez5yEdWPX6S8iEi8i/YA/AHOMMV6sy0T/JyIN7Zu292A1iuFsA1qKSHwFb2suVnD6C6W9AbBuNk8UkY72e00SkWH2ug+BU0VkvIgk2GXq6Thua/+NbGPMZmAh8KiIJIpICnAD4e8FVOQ54HdYPZlgDYEDwB4RaQHc61i3CPgNeExETrDL0qeaZQjJGPMb8CnwlIg0EhGPfQN7QE0eR9UMDQSquvZgNfwrReQA1s3SeVg3DAHeAJZj3Qv4lPKNPMC/7HXr7elvjnVbgd1YZ9GzsO4n/GSvuxOrx7Ee67r4v4AZEcr6JbAK2CoiO8Nlsi9F+YPBLEf6POBx4G37MssPwEX2uv1YjfEldpl/BjLsTefYr/ki4u9djMC6QbsFq74eMsZ8FqHsYRljdhn7qa0Qq/8CdAP2Ah9h9db823nt8p4JbMK6eX1VdcpQgeuwLuP9iPV/+S7QLArHUUdIQn+GlIouEcnFuun4eYh16Vg3WVvWcrGUciXtESillMtpIFBKKZfTS0NKKeVyUesRiMgMEdkuIj+EWS8i8pz9NfsVItItWmVRSikVXjSfI54JPE/oZ5zBeuriLHvqCUy1XyNq2rSpad26dc2U8Chy8OBBTjgheMwz5af1E5nWT2RaP7BkyZKd9pc/y4laIDDGZEvkIYkvBV63H337VkROFJFm9vPHYbVu3ZrFixfXZFGPCllZWaSnp9d1MY5aWj+Raf1EpvUDIrIx3Lq6vFncgrJfp8+j7FftlVJK1YJoXhqqSKjRDEPeuRaRm4GbAZKTk8nKyopiserGgQMHjsv3VVO0fiLT+olM6yeyugwEeZQda6YlZceaCTDGvAS8BJCWlmaOxy6edl0j0/qJTOsnMq2fyOoyELwP3CEib2PdJN5b0f0BpY4mxcXF5OXlUVBQUNdFISkpidWrV9d1MY5abqqfxMREWrZsSVxcXKW3iVogEJG3sMaVbyrWL1M9hDUsLcaYacDHWGPQr8Ma23106D0pdXTKy8ujYcOGtG7dmvC/21I79u/fT8OGDSvO6FJuqR9jDPn5+eTl5dGmTZtKbxfNp4ZGVLDeALdH6/hKRVtBQcFREQSU8hMRmjRpwo4dO6q0nQ4xodQR0CCgjjbV+UxqIFBKKZfTQKDUMSo/P5/U1FRSU1M588wzadGiRWC5qKgo4raLFy9m3LhxFR6jd+/eNVLWrKwskpKSAuVLTU3l88/LjUBe63Jzc+nUqVPIdfPnz6djx46kpqZy+PDhkHmOF3X51JBS7vHjFGjSHZIzStO2ZUJ+DnS4r1q7bNKkCcuWLQNg4sSJNGnShAkTJgTWl5SUEBsb+k88LS2NtLS0Co+xcOHCapUtlH79+vHhhx+GXR/4IXWPJ+RyOF6vl5iYmCqXx+v1Rlw/a9YsJkyYwOjRx/9zLNojUKo2NOkOC660Gn+wXhdcaaXXoFGjRnHPPfeQkZHB/fffz6JFi+jduzddu3ald+/erFmzBrDO0P/whz8AMHnyZMaMGUN6ejpt27blueeeC+yvQYMGgfzp6ekMHTqUdu3acc011+Afufjjjz+mXbt29O3bl3HjxgX2Wxm5ubm0b9+e2267jW7dujF//vwyy5s3b+bee++lU6dOdO7cmXfeeSdQnoyMDK6++mo6d+5c6eOF2q6kpITrr7+elJQUhg4dyqFDh5g+fTqzZ8/m4Ycf5pprrqn0/o9V2iNQqiYsGQ+7l0XOU685fDkY6jWDw79BUntY+RdrCqVxKpz7TJWLsnbtWj7//HNiYmLYt28f2dnZxMbG8vnnnzNp0iTmzp1bbpuffvqJzMxM9u/fzznnnMOtt95a7jn0pUuXsmrVKpo3b06fPn34+uuvSUtL45ZbbiE7O5s2bdowYkT4hwXnz59PampqYHnu3LnExMSwZs0aXn31VV588UVyc3PLLM+dO5dly5axfPlydu7cSffu3enfvz8AixYt4ocffqjSY5LB2/mP98orr9CnTx/GjBnDiy++yIQJE1iwYAF/+MMfGDp0aJX2fyzSHoFStSW+sRUEDm2yXuMbR+Uww4YNC1wq2bt3L8OGDaNTp07cfffdrFq1KuQ2v//970lISKBp06accsopbNu2rVyeHj160LJlSzweD6mpqeTm5vLTTz/Rtm3bQGMcKRD069ePZcuWBaYzzjgDgNNPP51evXoF8jmXFyxYwIgRI4iJiSE5OZkBAwaQk5MTKE9Vg0Co7Vq1akWfPn0AGDlyJAsWLKjyPo912iNQqiZU5szdfzmo04Pw81To/FDZewY1xDnc8oMPPkhGRgbz5s0jNzc37DALCQkJgfmYmBhKSkoqlacmftgqeHho53Kk/YcbVnrevHn85S9WL2v69Onl7oUEbxf8uKUbHwnWHoFStcEfBPrOhpSHrVfnPYMo2bt3Ly1aWIP6zpw5s8b3365dO9avX09ubi5A4Bp+Tenfvz/vvPMOXq+XHTt2kJ2dTY8ePSJuc/nllwd6HZW5Ib5p0ya++eYbAN566y369u1bI2U/lmggUKo25OdYjb+/B5CcYS3n50T1sPfddx8TJ06kT58+FT4lUx316tXjxRdf5MILL6Rv374kJyeTlJQUMq//HoF/evfddyvc/+WXX05KSgpdunRh4MCBTJkyhVNPPbVG30P79u157bXXSElJYdeuXdx66601uv9jwTH3m8VpaWlGf5jGfY7G+lm9ejXt27ev62IAdTuWzoEDB2jQoAHGGG6//XbOOuss7r777jopSzhuGWvIL9RnU0SWGGNCdpG0R6CUOiIvv/wyqampdOzYkb1793LLLbfUdZFUFenNYqXUEbn77ruPuh6AqhrtESillMtpIFBKKZfTQKCUUi6ngUAppVxOA4FSx7CtW7cyfPhwUlJS6NChAxdffDFr166t62KVM3PmTO644w4Apk2bxuuvv14uT6QhoZ15/vWvfwWWKzucdmW0bt2azp07B77nUFP7PVLp6emEemS+sLCQQYMGkZqaesRf5NOnhpSqBVOmQPfukOEYUSIzE3Jy4L7qjUKNMYbLL7+c66+/npdffpmGDRuybNkytm3bxtlnnx3IV91hmqNl7Nix1d7WHwiuvvpqoPLDaVdWZmYmTZs2Dbs+eGjvSEN9O0VjqOylS5dSXFwcGIr8SGiPQKla0L07XHml1fiD9XrllVZ6dWVmZhIXF1emYU1NTaVfv37lhlsuKChg9OjRdO7cma5du5JpF2TVqlX06NGD1NRUUlJS+Pnnnzl48CC///3v6dKlC506dSp3tunz+WjdujV79uwJpJ155pls27aNDz74gJ49e9K1a1cGDRoUcvC6yZMn8+STTwKwZMkSunTpwnnnnccLL7wQyJObm0u/fv3o1q0b3bp1C/wuwgMPPBD4hvLTTz9dZjjtXbt2cdlll5GSkkKvXr1YsWJF4Hi33XZbyGG2KyM9PZ1JkyYxYMAAnn322XLLX3zxBV27dqVz586MGTOGwsJCwOphPPzww/Tt25c5c+ZU+nihtnvzzTfp3bs3nTp1YtGiRWzfvp2RI0eybNkyUlNT+eWXX6r0noJpj0CpGjB+PFR0Yta8OQweDM2awW+/Qfv28Je/WFMoqanwzDPh9/fDDz9w7rnnhl3vHG75qaeeAmDlypX89NNPXHDBBaxdu5Zp06Zx1113cc0111BUVITX6+Xjjz+mefPmfPTRR4A1XpGTx+Ph0ksvZd68eYwePZrvvvuO1q1bk5ycTN++ffn2228REaZPn86UKVMCxw5l9OjR/OMf/2DAgAHce++9gfRTTjmFzz77jMTERH7++WdGjBjB4sWLeeyxx3jyyScDP3CTlZUV2Oahhx6ia9euvPfee3z55Zdcd911gbPltWvXkp2dHXGYbYCMjIzAmfv1118f+H7Enj17+OqrrwD44IMPAssFBQWcddZZfPHFF5x99tlcd911TJ06lfHjxwOQmJhYrdFMndtNmzaNgwcPsnDhQrKzsxkzZgw//PAD06dPL1MXR0J7BErVksaNrSCwaZP12jg6o1AHOIdbXrBgAddeey1gDRR3+umns3btWs477zweeeQRHn/8cTZu3Ei9evXo3Lkzn3/+Offffz/z588POXbQVVddFegpvP3221x11VUA5OXlMXjwYDp37swTTzwRdthrsALMnj17GDBgAECgfADFxcXcdNNNdO7cmWHDhvHjjz9W+H6d73HgwIHk5+cHgtjgwYMrHGYbrF6Wf8A655fk/O8veHnNmjW0adMmcCnu+uuvJzs7O+x2lRW8nX947/79+7Nv374yvbGaoD0CpWpApDN3P//loAcfhKlT4aGHyt4zqKqOHTtGHLitMsM5X3311fTs2ZOPPvqIwYMHM336dAYOHMiSJUv4+OOPmThxIhdccAGDBw8ODB3x8MMPc8kll7Bu3Tp27NjBe++9x5///GcA7rzzTu655x6GDBlCVlYWkydPDls+Y0zYIZ+ffvppkpOTWb58OT6fj8TExIqqI+R79O+/MsNsRxJuqOyKxmoLNVS21+sN9OSGDBnCww8/XOF20R4qW3sEStUCfxCYPRsefth6dd4zqI6BAwdSWFjIyy+/HEjLyckJXMJw6t+/P7NmzQKsyySbNm3inHPOYf369bRt25Zx48YxZMgQVqxYwZYtW6hfvz4jR45kwoQJfP/99/Ts2TNwpjxkyBBEhMsvv5x77rmH9u3b06RJE6DssNevvfZaxPKfeOKJJCUlBS6B+Mvn30+zZs3weDy88cYbgZumDRs2ZP/+/SH353yPWVlZNG3alEaNGlWqLqurXbt25Obmsm7dOgDeeOONQA8nnJiYmEBdhgoCofh7XwsWLCApKSnsCK/VpT0CpWpBTo7V+Pt7ABkZ1nJOTvV7BSLCvHnzGD9+PI888gj169endevWPPPMM/z6669l8t52222MHTuWzp07Exsby8yZM0lISOCdd97hzTffJC4ujlNPPZX//d//JScnh3vvvRePx0NcXBxTp04NefyrrrqK7t27l/mdg8mTJzNs2DBatGhBr1692LBhQ8T38OqrrzJmzBjq16/P4MGDy5T3iiuuYM6cOWRkZATOkFNSUoiNjaVLly6MGjWKrl27ljn26NGjSUlJoX79+hUGolCc9whSUlJCPubqlJiYyKuvvsqwYcMoKSmhe/fuR/RUVDiNGzemd+/e7Nu3jxkzZtT4/nUY6qPE0TjM8tHkaKwfHYb62OG2+tFhqJVSSlWJBgKllHI5DQRKHYFj7dKqOv5V5zOpgUCpakpMTCQ/P1+DgTpqGGPIz8+v1OO2TvrUkFLV1LJlS/Ly8tixY0ddF4WCgoIq//G7iZvqJzExkZYtW1ZpGw0ESlVTXFxc4Ju7dS0rK6vMo5SqLK2fyPTSkFJKuZwGAqWUcrmoBgIRuVBE1ojIOhF5IMT6JBH5QESWi8gqERkdzfIopZQqL2qBQERigBeAi4AOwAgR6RCU7XbgR2NMFyAdeEpE4qNVJqWUUuVFs0fQA1hnjFlvjCkC3gYuDcpjgIZiDaXXANgFVG1YQKWUUkckmoGgBbDZsZxnpzk9D7QHtgArgbuMMb4olkkppVSQaD4+GmrA7OBv3gwGlgEDgTOAz0RkvjFmX5kdidwM3AyQnJxc5leJjhcHDhw4Lt9XTdH6iUzrJzKtn8iiGQjygFaO5ZZYZ/5Oo4HHjPXVzHUisgFoByxyZjLGvAS8BNboo0fbKJQ14WgcXfNoovUTmdZPZFo/kUXz0lAOcJaItLFvAA8H3g/Kswk4H0BEkoFzgPVRLJNSSqkgUesRGGNKROQO4BMgBphhjFklImPt9dOAvwIzRWQl1qWk+40xO6NVJqWUUuVFdYgJY8zHwMdBadMc81uAC6JZBqWUUpHpN4uVUsrlNBAopZTLaSBQSimX00CglFIup4FAKaVcTgOBUkq5nAYCpZRyOQ0ESinlchoIlFLK5TQQKKWUy2kgUEopl9NAoJRSLqeBQCmlXE4DgVJKuZwGAqWUcjkNBEop5XIaCJRSyuU0ECillMtpIFBKKZfTQKCUUi6ngUAppVxOA4FSSrmcBgKllHI5DQRKKeVyGgiUUsrlNBAopZTLaSBQSimX00CglFIup4FAKaVcTgOBUkq5nAYCpZRyOQ0ESinlchoIlFLK5TQQKKWUy0U1EIjIhSKyRkTWicgDYfKki8gyEVklIl9FszxKKaXKi43WjkUkBngB+B2QB+SIyPvGmB8deU4EXgQuNMZsEpFTolUepZRSoUWzR9ADWGeMWW+MKQLeBi4NynM18G9jzCYAY8z2KJZHKaVUCNEMBC2AzY7lPDvN6WygsYhkicgSEbkuiuVRSikVQtQuDQESIs2EOP65wPlAPeAbEfnWGLO2zI5EbgZuBkhOTiYrK6vmS1vHDhw4cFy+r5qi9ROZ1k9kWj+RRTMQ5AGtHMstgS0h8uw0xhwEDopINtAFKBMIjDEvAS8BpKWlmfT09GiVuc5kZWVxPL6vmqL1E5nWT2RaP5FF89JQDnCWiLQRkXhgOPB+UJ7/B/QTkVgRqQ/0BFZHsUxKKaWCRK1HYIwpEZE7gE+AGGCGMWaViIy1108zxqwWkf8CKwAfMN0Y80O0yqSUUqq8aF4awhjzMfBxUNq0oOUngCeiWQ6llFLh6TeLlVLK5TQQKKWUy2kgUEopl9NAoJRSLqeBQCmlXK7CQCAiHhHpXRuFUUopVfsqDATGGB/wVC2URSmlVB2o7KWhT0XkChEJNX6QUkqpY1hlv1B2D3AC4BWRw1gDyhljTKOolUwppVStqFQgMMY0jHZBlFJK1Y1KDzEhIkOA/vZiljHmw+gUSSmlVG2q1D0CEXkMuAv40Z7ustOUUkod4yrbI7gYSLWfIEJEXgOWAiF/kF4ppdSxoypfKDvRMZ9Uw+VQSilVRyrbI3gEWCoimVhPDPUHJkatVEoppWpNhYFARDxYPxrTC+iOFQjuN8ZsjXLZlFJK1YIKA4ExxicidxhjZlP+pyaVUkod4yp7j+AzEZkgIq1E5CT/FNWSKaWUqhWVvUcwxn693ZFmgLY1WxyllFK1rbL3CB4wxrxTC+VRSilVyyo7+ujtFeVTSil1bNJ7BEop5XJ6j0AppVyusqOPtol2QZRSStWNiJeGROQ+x/ywoHWPRKtQSimlak9F9wiGO+aDh5S4sIbLopRSqg5UFAgkzHyoZaWUUsegigKBCTMfalkppdQxqKKbxV1EZB/W2X89ex57OTGqJVNKKVUrIgYCY0xMbRVEKaVU3ajKD9MopZQ6DmkgUEopl9NAoJRSLqeBQCmlXE4DgVJKuVxUA4GIXCgia0RknYg8ECFfdxHxisjQaJZHKaVUeVELBCISA7wAXAR0AEaISIcw+R4HPolWWZRSSoUXzR5BD2CdMWa9MaYIeBu4NES+O4G5wPYolkUppVQYlf09gupoAWx2LOcBPZ0ZRKQFcDkwEOgebkcicjNwM0BycjJZWVk1XdY6d+DAgePyfdUUrZ/ItH4i0/qJLJqBINSgdMHjEz0D3G+M8YqEH8POGPMS8BJAWlqaSU9Pr6EiHj2ysrI4Ht9XTdH6iUzrJzKtn8iiGQjygFaO5ZbAlqA8acDbdhBoClwsIiXGmPeiWC6llFIO0QwEOcBZItIG+BXrtw2udmZw/vKZiMwEPtQgoJRStStqgcAYUyIid2A9DRQDzDDGrBKRsfb6adE6tlJKqcqLZo8AY8zHwMdBaSEDgDFmVDTLopRSKjT9ZrFSSrmcBgKllHI5DQRKKeVyGgiUUsrlNBAopZTLaSBQSimX00CglFIup4FAKaVcTgOBUkq5nAYCpZRyOQ0ESinlchoIlFLK5TQQKKWUy2kgUEopl9NAoJRSLqeBQCmlXE4DgVJKuZwGAqWUcjkNBEop5XIaCJRSyuU0ECillMtpIFBKKZfTQKCUUi6ngUAppVxOA4FSSrmcBgKllHI5DQRKKeVyGgiUUsrlNBAopZTLaSBQSimX00CglFIud9wHgilTIDOzbFpmppVeXcaYwOQzPnzGh9fnDUwlvpLAVOwtpthbzKOPefn082KKvEWB6bPPS3jscWsbAJ/xYYw5gnerlFJVF1vXBYi27t3hyith9mzo0ecgH35ykNtGN+GFV3eycU8BYDfsmMC8nw9f6byvdB5xHMDZbtvpxhjEXjAYRIQW7eoxfHgznpm+lV59D/PtgnqMv7EZz0zfwi+7D1NYUsi6/HXWPgx4PB48dpwWETziKTMB5dKc6SKCIIFX/37+/mQsaWmGjAwCaVmZwuLFwr33mkBa6VtyzIvzjSuljhfHfSDIyLCCwBVXQFqPOL5e0JQR1+1n/bpYNvzSMOQ2gYazTKJwpCfrV1xVwG0jW9CzTxHfLYznxlsPUHzoBL7LrM9vufspMCcRH2+ITzDEJ/hK5+N9xCX4iI/34RVvIFgZTNh5DIGg4g9GGGgZCEi/0bPvIb5bUL9MQLJ34KyMUo50j6e0M+lxdCyDg4h/2RmQPOLhxWca0KVbEf0GlATyL5yfwLIlcYy7p6Dc9l7jZffh3WWCXfAxQwUtZxAMlefJJzx0T4P0DBPI4w+M991HyH2o6JoyxTqB85+sgNWLz8kh8H/ipnLUhqgGAhG5EHgWiAGmG2MeC1p/DXC/vXgAuNUYs7ymy5GRAQMHwty58QC8MjWppg9RJV9+mgjAs080cqQ2qdS2pcHBEJ8AiQn+wGFISMB6jTckJFrr4+MNCQnWFG+v/92Fhdx6TQvSehWx+Lt4Rlx3iK2bGvLB2w2JiTHExGBNsdZ8bKzB4ymb5s/niTGl28QaPB5DbKyV7pwP5PcYYmJLOLvjYcaOOpl/vLKd3v0L+GZ+IneOSeK5GdvYV3i4zHs2GEp8JeQfzi8b7Cjb+wq08aZ0u0Dj7Q9iQb22lu3qcVWYntq6/MOBbfz5RazJ2VsDygUoZ68s1LpAHnufz/09kW7nehmQUdrzzM6KYcliD+P/WFTh58JnfOwr3FepS4vOHrDTc39PoOu53kBw9hkf87+KZemSWMbdU1BuO+fJh7McVVkO3ocxhpbtEhg67BT+MWM75/Xzfzas5Q27Cyp8f6EUeYvI3Z0LVD6gt2qfwNBhJ/P8qzvo3a+Qb+Yncvvoprz4aj6b9xaWORmBoJMTx0lMReucJ0nBeYHAZ6N/upe4mDgSYxNrPCBJtK5Ji0gMsBb4HZAH5AAjjDE/OvL0BlYbY3aLyEXAZGNMz0j7TUtLM4sXL65SWTIzrctDw4YX8NasOB5/di/de1X8xxXKkZwU5nwbz33jTmToiEO8+1Z9Hnp0Lx06FVNUKKxZkUvz1m0pLIKiQqGwQCgqEgoLobBQKPJPRVBQUDpfWCAUFlnLVj57+yKhsACK/OsKhEI7v893tJ3ZWp9BfwDyeACxgo/HY9W58XmJjYtB7HQRAusRK8AE0sV6FY+V7szrTxf/NgIHDggbN8RySrKX7dtj6JxSTHIzL3HxduCNN8THQ1y8IS7OCrpx8VbAjfUvO15j430kxENsnNWri7ODd2ysL5AnLsFnvcZZ67/7OtEOQL8FBSRr2R/UjAFjwOezJ6/g88GG5Rs5rVNrvF4wPsH4rHSvj8C8M7/P+NNL161YksiTf23C+En5dOhcxMplCTz32Enc/sfdnNOxCJ9X8HrB67X24w3sT+w0f7oE8vjXld+GwP6c23vtbbbkxbAgK5Ezzi7ml7Vx9BlQwKnNvY7/e4Pg/L8sTafMMogY8rfu5OTmTQOfh1Dbicc6T3CmbVgfy5xZDehxXiGLvklgxHX7OfOcksDnM7CtWMfx71fE+bmzL7k6Pov+V8Qg9ucwkIYJvA/xGFavTOCZR5syftJ2LrkEDmxtFrjc7eytVERElhhj0kKui2IgOA+rYR9sL08EMMY8GiZ/Y+AHY0yLSPutaiDwB4HZsyG1127m/WcPE25uxTPTf+O8fqVnF+GuiweVsdz6ylxPF4SF2QncNroJU2fuonf/QhZmJ3DrqJOY9tpu+vQvYlXOKjp271jp93UksjPjuW10Y64aeYh33jyBKc/uoWv30j/0khLrj7OkpPQP2D/vdTQG3hLHvD+9xJq39uGY90JJSWlD4N/f/MwEvluYQI9ehfToU4Tx+Rs6KdPg5W/dReOTTypNtxtEY683prRB86f791M+L4EG0J/+y7pYNuXG0rxFCSef4qOo2AqqxcWlQbeoSCgusgJuTYuNNcTEGooKhbh4KCqChASrFxJoxH0chUG8+jye0t6nx9+r9JTOHzokHDzg4YQGPho0MI7/O///m2AgdLoJkW6O/bpL61lI7i8JVQ4CEDkQRPPSUAtgs2M5D4h0tn8D8J9QK0TkZuBmgOTkZLKysipdiLffbsWkSfsR2cOy7wytGxv+NGk3C/7TkBZJm8pvYJyzoYNkuPRI2372QWvun7iRE+vt5sccSKpnuH/iSXzyXiMaJeZSeKiQld+ttDKLY1/VmQ9DEJYta8wjf+3CpAe/JzV1N21Oa8yEO7ow6cHlpKbuDuRzbIQHiAdrxgPEhT9GVSxb2pifVqVwzcj1fPhBS4YOXUNq190h8xYcLCDxhMSaOXCIcvzfX0vLMf7uFWHLAVaj4vUKxcVCcbGHkhKP9VrsoahYKHGkBecpu+yYL/JQUiKsWNGYtWuSaNduD50677HOEmMMHvts0+M/E/U40jwGb3Ex8Ymx1voYHD0n4+gtmcAZapn1Mf4zUivti89PJSuzGQPP38KFF22xGmZP6bYejwlc+ovxp8WYwLpQaf5Gv3S5sv8neXz4QUv+OCHy/0lF/J8ff2AAKRMgnGmY0pOHlSsa8/enOnDB737j00+bcdf41XTouBefceSzA5JxnKQ49+FPN3ZauXX2toF1Psrsz1++zz9rRlZmM669NheRXKrQDFYomj2CYcBgY8yN9vK1QA9jzJ0h8mYALwJ9jTH5kfZbnUtDRztjDF999RX9B/Sv1I3giub9j6EaTLn555+uT5duRZzXr/Sa78LsBJYvjeOWcfsC2zvL5leVp6ic1+dDPUX17fx6ES+FlKkfDLnLc2mT2iZiHVpFiXzW5w/Q/rKVKUe/w+XKFeqGu/OegfVPyvUWncvOJ7iC05x5v86OZ+z1jbnuhkO8/kr9QG8x0vv1v6fVi1fTPq19uXsoofIGpzvTFmYnMO6GU7lm9F5mvZrEc69stT4rIZ6qK3OyU8EDBuE+D6HyrprzJTf+aRRPz9gReKDh7jEnM/2RmXQYmlHm/7jMfaDgfTvWb1i2gTapbcqtD/64ONc7y+H/bPjL0XHYQCKp6HNYFY03vMyXPw5i7D0Due6GQ8ya0YjZ05aS0f4z6FD5mwR11SPIA1o5llsCW4IziUgKMB24qKIgcLwqc8Mxyr3Xvz5YPu2qS6wJGldrn9VpfP6dPZ93pm5lwCVdAGh9CTSbupQl2ftpfUnfcsfYErOFNieGDwR+lb0R6P9DnTc/m9nTtpMxJBWAtkMMzacuY3H2Plpf0jcQLIMDrrNh9H+XBAjMBwdgZx6vsZ788gdWn8/Hqne/ZOykUTw9Ywu9+h2m63n1GHvtyUz/v5l0vHKgvyIDgp/aMsZQVFJULtj468T5xFZgu6D5xf/6iPETr+GVN/bTd4CX8wce4MaRTXn98X9x3jWXhd0uuO6rcuk0VNrcXwcy567h9O9xO6ZxOmf0yKLlXXeyKG8Kv2/cGqfgE9ngnrh//a+eX2nVqFXY9aG2n5OXzuxxwznv3JvwNujP0HOzaTHuZb7Z9CiDGjQrt100TqqNMXy97lJuv+Mc3njpOzKu6MqlPZZy5ZhWzJ7xOzI61MxxohkIcoCzRKQN8CswHLjamUFETgP+DVxrjFkbxbIovx+nQJPukOy4wLgtE/JzqnR24RTqj7qigPbA/QILLoads62ybMvkd54r+d39syEmvvwxEOJiauialMP99wELLoQdpeUY5LmSQWHKES3zfvUw567hpPe6E9M4nTN7WY1fzpYnuOykMwP5wgW6TTGbaHtS2yMqw6rNA5gz7ioyzr0LTsjgsm6ZJI17lpyNT3FxvQgnCcaAKQFfEfiK7Vd78haVXS43lc8/8eoi2NcE5l8GjbvC7u8ZOOgyBtafASugtJvhb3hN+bRAr9Zad+a+LdRb9p4jzX41JvT2GP40BDjsha+HQlIH2Psj56cP4vx6z8Iy/5v392yc/y8h5o9g/aqFGcz+6/NkyBT4bggZfMHsGZ+S80tXqnibIKyoXRoCEJGLgWewHh+dYYz5PxEZC2CMmSYi04ErgI32JiXhui5+x+yloQoa4KysLNLT06Nfjm2ZsOBK6Fva8JVZPlI+bwV/+I7GIX8RrPobnDoYfvsvnH0nNGpnNyol9qsXTAnrflnLma1PA+MtXWdKyi777GXn9qGWg7cv3AUHf4H4plC0ExqcCfEn2Y+WxACe0vlwrziXnfOx5ffjiQm934MbYdNsaNLDqptWQ+GE0+1GzWe9Gp/9nvxp1t3uX7fk0aJZcmmeMvmrsFy4E/auhHrN4PAWOKEtxCSWb7D9jbix56NKrHoMTgt+leC00rwlxhDrf6QnTJ5y+/Tn9R4GcxikHsTUK92kTCCigvngbZzrK7m9rxgogU4PQsrD5fddgTp5aihaqhwIonAGXC0VNMDVCgTGgK8QvAX2dBhKDtvzh0rnSw7Zr4etPHtXwea3ICkV9iyF5Asg4eTyjbUpBm+h9RrqLM44XwvtD6qvolLXILtBlRjA2eDGlqZLjJ3Pvz42aNmeL9gERZsgvhUktLLeh/FinTX6AK/9R+x8dTSggXRnw+rf3vnqWO9Mi8i+Sy9CIHhQdr7IZ4iPibPSxX9X3zHv3zbsvCNv8VYo3gZxLSCxDXjiQMJMnnir/jzx9nKcYz6+dF4c6TF2ekyCPcVDbKJjOQH2LoblN8HpN8DGV+Dc16Bp/2p/UrJyckjv3r3qG+7MhiXXH1k5/M/8Homd2bB0NLS5CXKnV+vEra7uERwdmnQP3wDXBOOzGktfQWmD7G9wA/N2epsb4ash0LQP7FgALYdB3qew8X3O3JMLC2baDfvhso27t6B0/z77WD57/kjsWgjEwLYvQvyBBy0TC55GVnpMqMYgzm5c40DiyzYMEtwg2NscXgtb/gFNLoL8/0DrP0PDc62GQWLAExuYn79mHf06draPHWutK/NANuVfxXFmJ1I+zf+6PQu+GwntJ8EvL0HXh+DkAUH/zyb0fFXWBS87b7gbn1WO5TdCq1Gw+TVIfQVOSQ/+XyPUIzcLc3JI79GjfN5Ql5Iipe34ChZdC+c8ABumQ+qfrTKEq8tw9XoktmXCipuh3xzrb/a0C4+81+rxQMPQIwlELMf3o2q2HNWxLROWjSktR8sLarwcx38gSM6wKmz+UDi5n9XonXYtbP0Ofs0qbVy9h60G3d8AOxveMg1woX1WXGhNphrd4q2fWK8bZ1qvkkgysbClHngSQBLshjPRfj0RYhMgLsGx3j8llp2PSbTWS7ydVs8+M7PXSwIcXgnr74fkkbBtFpzzPJzU1/ENLYK+gSXl10VqECqzvC0Tvv4jpL8XoodU/maxN3cntDyj6nVdkW2ZsGgk9LP/qJoPqrs/9pU3Q/93reO2+UPVyuHxQIMGR16GnGtL6yIKDU6l5OeUPab/bzg/R8sRpXIc/4EArMpKzoDNc63l9VPLrhd/Yxof1Ij6lxtAbJPSdH9DXaZRtru2/mWJd+zHbtAPrYbNj0HT/4H89+DMp6Fxf/B4+HrtWtI7dAjdGAc3ysETVD59WyYsmAQD5toN8OV188e+a7Fr/siOmXIcDWWA0Jds/X/DtclF5XBHINiWCduy4LQb4Nd34fS/QlIfu7GOKx3XwHnm62x8Q81HaphDNcLbs2DhU5A+z26Ar7Ua4Bb2H96GDdAi4peqa4b+sWs5juYyqDpx/AcC/yWHfnOsa52/DYdvRsDpb8OpA2vmmmZl7F6iDbBS6qh0/AeC4DPg5oNKG+Bm59deObQBVkodpY7/QKANsFJKRVTB0E9KKaWOdxoIlFLK5TQQKKWUy2kgUEopl9NAoJRSLqeBQCmlXE4DgVJKuZwGAqWUcjkNBEop5XIaCJRSyuU0ECillMtpIFBKKZfTQKCUUi6ngUAppVxOA4FSSrmcBgKllHI5DQRKKeVyGgiUUsrlNBAopZTLaSBQSimX00CglFIup4FAKaVcTgOBUkq5nAYCpZRyOQ0ESinlchoIlFLK5TQQKKWUy0U1EIjIhSKyRkTWicgDIdaLiDxnr18hIt2iWR6llFLlRS0QiEgM8AJwEdABGCEiHYKyXQScZU83A1OjVR6llFKhRbNH0ANYZ4xZb4wpAt4GLg3KcynwurF8C5woIs2iWCallFJBYqO47xbAZsdyHtCzEnlaAL85M4nIzVg9BoADIrKmZot6VGgK7KzrQhzFtH4i0/qJTOsHTg+3IpqBQEKkmWrkwRjzEvBSTRTqaCUii40xaXVdjqOV1k9kWj+Raf1EFs1LQ3lAK8dyS2BLNfIopZSKomgGghzgLBFpIyLxwHDg/aA87wPX2U8P9QL2GmN+C96RUkqp6InapSFjTImI3AF8AsQAM4wxq0RkrL1+GvAxcDGwDjgEjI5WeY4Bx/Wlrxqg9ROZ1k9kWj8RiDHlLskrpZRyEf1msVJKuZwGAqWUcjkNBFEkIrkislJElonIYjvtJBH5TER+tl8bO/JPtIfbWCMigx3p59r7WWcPyRHqsdujnojMEJHtIvKDI63G6kNEEkTkHTv9OxFpXatv8AiFqZ/JIvKr/RlaJiIXO9a5rX5aiUimiKwWkVUicpedrp+hI2WM0SlKE5ALNA1KmwI8YM8/ADxuz3cAlgMJQBvgFyDGXrcIOA/rexf/AS6q6/dWzfroD3QDfohGfQC3AdPs+eHAO3X9nmugfiYDE0LkdWP9NAO62fMNgbV2Pehn6Agn7RHUvkuB1+z514DLHOlvG2MKjTEbsJ6k6mEPudHIGPONsT6drzu2OaYYY7KBXUHJNVkfzn29C5x/LPWewtRPOG6sn9+MMd/b8/uB1VgjEehn6AhpIIguA3wqIkvsYTIAko39XQn79RQ7PdxwGy3s+eD040VN1kdgG2NMCbAXaBK1kteeO+zReWc4Lnu4un7sSzZdge/Qz9AR00AQXX2MMd2wRlm9XUT6R8gbbriNSg3DcRyqTn0cj3U1FTgDSMUag+spO9219SMiDYC5wHhjzL5IWUOkuaKOqkoDQRQZY7bYr9uBeVgjsm7zj7Bqv263s4cbbiPPng9OP17UZH0EthGRWCCJyl9qOSoZY7YZY7zGGB/wMtZnCFxaPyIShxUEZhlj/m0n62foCGkgiBIROUFEGvrngQuAH7CG1bjeznY98P/s+feB4fZTC22wfqNhkd3V3S8ivexrldc5tjke1GR9OPc1FPjSvgZ8zJKyw7JfjvUZAhfWj/1+XgFWG2P+7liln6EjVdd3q4/XCWiL9cTCcmAV8Cc7vQnwBfCz/XqSY5s/YT3ZsAbHk0FAGlYD8AvwPPY3wo+1CXgL6/JGMdaZ1w01WR9AIjAH66bgIqBtXb/nGqifN4CVwAqsRqqZi+unL9ZlmhXAMnu6WD9DRz7pEBNKKeVyemlIKaVcTgOBUkq5nAYCpZRyOQ0ESinlchoIlFLK5TQQqKOSiDRxjLi5NWgEzvgKtk0TkecqcYyFNVfiuicio0Tk+bouhzr2RO2nKpU6EsaYfKxhFRCRycABY8yT/vUiEmussWBCbbsYWFyJY/SukcIqdYzTHoE6ZojITBH5u4hkAo+LSA8RWSgiS+3Xc+x86SLyoT0/2R6sLUtE1ovIOMf+DjjyZ4nIuyLyk4jMcoxPf7GdtsAet/7DEOWKEZEnRCTHHhzuFjv9HhGZYc93FpEfRKR+hHKPEpH3ROQDEdkgInfY+1gqIt+KyEl2viwRecbe9gcR6RGiTCeLyFy7TDki0sdOH+DoWS31f/tduZv2CNSx5mxgkDHGKyKNgP7GmBIRGQQ8AlwRYpt2QAbWGPZrRGSqMaY4KE9XoCPWmDNfA33E+jGhf9rH2CAib4Up0w3AXmNMdxFJAL4WkU+BZ4AsEbkc6xuutxhjDonITxHK3ckuSyLWt1vvN8Z0FZGnsYZCeMbOd4IxprdYAxnOsLdzehZ42hizQEROAz4B2gMTgNuNMV+LNXhbQZj3pFxEA4E61swxxnjt+STgNRE5C2vogbgw23xkjCkECkVkO5BM2WGIwRqDJg9ARJYBrYEDwHpjjWUP1hAQN1PeBUCKiAx1lOssO3iMwhoS4Z/GmK8rUe5MY421v19E9gIf2OkrgRRHvrfA+g0DEWkkIicGlWkQ0EFKh9JvZJ/9fw38XURmAf/2v2flbhoI1LHmoGP+r1gN5+VijU+fFWabQse8l9Cf+1B5KvuDJALcaYz5JMS6s7ACSnNHWqRyO8vhcyz7gsodPDZM8LIHOM8Yczgo/TER+QhrjJ5vRWSQMeankO9KuYbeI1DHsiTgV3t+VBT2/xPQVkp/t/aqMPk+AW4Va4hkRORssUafTcK6RNMfaBLUYzjScl9lH6sv1mWpvUHrPwXu8C+ISKr9eoYxZqUx5nGsG+rtqnl8dRzRQKCOZVOAR0XkayCmpndun03fBvxXRBYA27B+sSrYdOBH4Huxfnj+n1hn708DLxpj1mLdR3hMRE6poXLvth9/nWbvO9g4IM2+ef0jMNZOH2/fYF4OHMb6vV7lcjr6qFIRiEgDY8wB+ymiF4CfjTFP13GZsrB+0L7CR2SVqgztESgV2U32zeNVWJd0/lm3xVGq5mmPQCmlXE57BEop5XIaCJRSyuU0ECillMtpIFBKKZfTQKCUUi73/wE5R1AJ9AuYQQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "my_learn_curve(title=\" Support Vector Machine - Final Model\", \n",
    "               estimator=census_pipeline.fit(X_c_train,y_c_train), \n",
    "               cv=stratified_kfold,\n",
    "               X=X_c_train, \n",
    "               y=y_c_train,  \n",
    "               label_suffix=\"rbf\", \n",
    "               color_=[\"orange\",\"blue\"],\n",
    "               shape = 'x-'  )\n",
    "\n",
    "plt.savefig(\"SVM_CENSUS_final_learning_curve.pdf\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "boolean-isolation",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/flufylobster/anaconda3/envs/ML39/lib/python3.9/site-packages/category_encoders/utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7823416538237339\n"
     ]
    },
    {
     "ename": "NameError",
     "evalue": "name 'y_c_test' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-30-d45c0d4e559a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maverage_precision_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_c_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcensus_pipeline\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_c_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_c_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_c_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maverage_precision_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_c_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcensus_pipeline\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_c_train\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_c_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_proba\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX_c_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'y_c_test' is not defined"
     ]
    }
   ],
   "source": [
    "\n",
    "print(average_precision_score(y_c_train, census_pipeline.fit(X_c_train, y_c_train).predict_proba(X_c_train)[:,1]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "yellow-queue",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/flufylobster/anaconda3/envs/ML39/lib/python3.9/site-packages/category_encoders/utils.py:21: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead\n",
      "  elif pd.api.types.is_categorical(cols):\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7479314579426083\n"
     ]
    }
   ],
   "source": [
    "print(average_precision_score(y_c_test, census_pipeline.fit(X_c_train, y_c_train).predict_proba(X_c_test)[:,1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "integral-reservation",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.96      0.80      0.87     19758\n",
      "           1       0.58      0.89      0.70      6271\n",
      "\n",
      "    accuracy                           0.82     26029\n",
      "   macro avg       0.77      0.84      0.79     26029\n",
      "weighted avg       0.87      0.82      0.83     26029\n",
      "\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "          -1       0.94      0.79      0.86      4940\n",
      "           1       0.55      0.84      0.67      1568\n",
      "\n",
      "    accuracy                           0.80      6508\n",
      "   macro avg       0.75      0.81      0.76      6508\n",
      "weighted avg       0.85      0.80      0.81      6508\n",
      "\n"
     ]
    }
   ],
   "source": [
    "X_c_test = test_data.drop('income', axis=1)\n",
    "y_c_test = test_data['income']\n",
    "print(classification_report(y_c_train, census_pipeline.fit(X_c_train,y_c_train).predict(X_c_train) ))\n",
    "print(classification_report(y_c_test, census_pipeline.fit(X_c_train,y_c_train).predict(X_c_test) ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "rubber-profit",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sklearn.metrics._plot.confusion_matrix.ConfusionMatrixDisplay at 0x7ff5a60f8550>"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(estimator=  census_pipeline.fit(X_c_train,y_c_train) ,\n",
    "                      X=X_c_train,\n",
    "                      y_true= y_c_train  , \n",
    "                      cmap=plt.cm.Blues,\n",
    "                       normalize=\"true\",\n",
    "                      display_labels = [\"Not Patient\" ,\"Patient\"]\n",
    "                        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "specific-alpha",
   "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": "3.9.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}