{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import warnings\n", "warnings.simplefilter('ignore', FutureWarning)" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import matplotlib.pyplot as plt\n", "import pandas as pd\n", "\n", "from sklearn import (\n", " ensemble,\n", " preprocessing,\n", " tree,\n", ")\n", "from sklearn.metrics import (\n", " auc,\n", " confusion_matrix,\n", " roc_auc_score,\n", " roc_curve,\n", ")\n", "from sklearn.model_selection import (\n", " train_test_split,\n", " StratifiedKFold,\n", ")\n", "from yellowbrick.classifier import (\n", " ConfusionMatrix,\n", " ROCAUC,\n", " PRCurve,\n", ")\n", "from yellowbrick.model_selection import (\n", " LearningCurve,\n", ")" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "plt.rcParams.update({'font.size': 16})" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Data\n", "\n", "[Vanderbilt Datasets](http://biostat.mc.vanderbilt.edu/wiki/Main/DataSets)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Gather data" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "X_train = pd.read_csv('data/X_train.csv')\n", "X_test = pd.read_csv('data/X_test.csv')\n", "y_train = pd.read_csv('data/y_train.csv')\n", "y_test = pd.read_csv('data/y_test.csv')" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
pclasssex_malesex_femaleagesibspparchfareembarked_Sembarked_Cembarked_Qembarked_
0-0.3573000.734868-0.734868-0.0590750.510668-0.434783-0.2299230.649414-0.501274-0.321694-0.045198
1-1.5535850.734868-0.7348681.602868-0.489262-0.434783-0.1204420.649414-0.501274-0.321694-0.045198
2-1.553585-1.3607881.3607880.469725-0.489262-0.4347832.031380-1.5398491.994917-0.321694-0.045198
\n", "
" ], "text/plain": [ " pclass sex_male sex_female age sibsp parch fare \\\n", "0 -0.357300 0.734868 -0.734868 -0.059075 0.510668 -0.434783 -0.229923 \n", "1 -1.553585 0.734868 -0.734868 1.602868 -0.489262 -0.434783 -0.120442 \n", "2 -1.553585 -1.360788 1.360788 0.469725 -0.489262 -0.434783 2.031380 \n", "\n", " embarked_S embarked_C embarked_Q embarked_ \n", "0 0.649414 -0.501274 -0.321694 -0.045198 \n", "1 0.649414 -0.501274 -0.321694 -0.045198 \n", "2 -1.539849 1.994917 -0.321694 -0.045198 " ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_train.head(3)" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
pclasssex_malesex_femaleagesibspparchfareembarked_Sembarked_Cembarked_Qembarked_
00.838984-1.3607881.360788-0.890046-0.489262-0.434783-0.4853800.649414-0.501274-0.321694-0.045198
10.8389840.734868-0.734868-0.314790-0.489262-0.434783-0.501655-1.5398491.994917-0.321694-0.045198
2-1.5535850.734868-0.734868-0.3612460.510668-0.4347831.152487-1.5398491.994917-0.321694-0.045198
\n", "
" ], "text/plain": [ " pclass sex_male sex_female age sibsp parch fare \\\n", "0 0.838984 -1.360788 1.360788 -0.890046 -0.489262 -0.434783 -0.485380 \n", "1 0.838984 0.734868 -0.734868 -0.314790 -0.489262 -0.434783 -0.501655 \n", "2 -1.553585 0.734868 -0.734868 -0.361246 0.510668 -0.434783 1.152487 \n", "\n", " embarked_S embarked_C embarked_Q embarked_ \n", "0 0.649414 -0.501274 -0.321694 -0.045198 \n", "1 -1.539849 1.994917 -0.321694 -0.045198 \n", "2 -1.539849 1.994917 -0.321694 -0.045198 " ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_test.head(3)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
survived
00
11
21
\n", "
" ], "text/plain": [ " survived\n", "0 0\n", "1 1\n", "2 1" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "y_train.head(3)" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
survived
01
10
21
\n", "
" ], "text/plain": [ " survived\n", "0 1\n", "1 0\n", "2 1" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "y_test.head(3)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Baseline model" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "from sklearn.dummy import DummyClassifier" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [], "source": [ "dummy = DummyClassifier(strategy='prior')" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.6219512195121951" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "dummy.fit(X_train, y_train)\n", "dummy.score(X_test, y_test)" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [], "source": [ "from sklearn.metrics import accuracy_score" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.6219512195121951" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" } ], "source": [ "accuracy_score(y_test, dummy.predict(X_test))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Evaluate model families" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [], "source": [ "X = pd.concat([X_train, X_test])\n", "y = pd.concat([y_train, y_test])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### `sklearn` does not like column vectors for the target" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [], "source": [ "y = y.values.ravel()\n", "y_train = y_train.values.ravel()\n", "y_test =y_test.values.ravel()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Classification" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [], "source": [ "# ! python3 -m pip install --quiet xgboost" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [], "source": [ "from sklearn import model_selection\n", "from sklearn.linear_model import (\n", " LogisticRegression,\n", ")\n", "from sklearn.tree import (\n", " DecisionTreeClassifier,\n", ")\n", "from sklearn.neighbors import (\n", " KNeighborsClassifier,\n", ")\n", "from sklearn.naive_bayes import (\n", " GaussianNB,\n", ")\n", "from sklearn.svm import (\n", " SVC,\n", ")\n", "from sklearn.ensemble import (\n", " RandomForestClassifier,\n", ")\n", "import xgboost" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [], "source": [ "models = [ \n", " DummyClassifier,\n", " LogisticRegression,\n", " DecisionTreeClassifier,\n", " KNeighborsClassifier,\n", " GaussianNB,\n", " SVC,\n", " RandomForestClassifier,\n", " xgboost.XGBRFClassifier,\n", "]" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "DummyClassifier AUC:0.478 STD: 0.06\n", "LogisticRegression AUC:0.843 STD: 0.04\n", "DecisionTreeClassifier AUC:0.753 STD: 0.04\n", "KNeighborsClassifier AUC:0.840 STD: 0.03\n", "GaussianNB AUC:0.824 STD: 0.04\n", "SVC AUC:0.848 STD: 0.04\n", "RandomForestClassifier AUC:0.845 STD: 0.03\n", "XGBRFClassifier AUC:0.866 STD: 0.04\n" ] } ], "source": [ "for model in models:\n", " cls = model()\n", " kfold = model_selection.KFold(\n", " n_splits=10, \n", " shuffle=True,\n", " random_state=123,\n", " )\n", " s = model_selection.cross_val_score(\n", " cls, X, y, scoring='roc_auc', cv=kfold,\n", " )\n", " print(\n", " f'{model.__name__:22} AUC:'\n", " f'{s.mean():.3f} STD: {s.std():.2f}'\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Stacking\n", "\n", "Source: https://miro.medium.com/max/2044/1*5O5_Men2op_sZsK6TTjD9g.png\n", "\n", "\"Stacking\"\n" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [], "source": [ "from sklearn.ensemble import StackingClassifier" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [], "source": [ "stack = StackingClassifier(estimators=[(m.__name__, m()) for m in models])" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "StackingClassifier AUC:0.867 STD: 0.04\n" ] } ], "source": [ "s = model_selection.cross_val_score(\n", " stack, X, y, scoring='roc_auc', cv=kfold,\n", ")\n", "print(\n", " f'{stack.__class__.__name__:22} AUC:'\n", " f'{s.mean():.3f} STD: {s.std():.2f}'\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create a model" ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "XGBRFClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n", " colsample_bytree=1, gamma=0, gpu_id=-1, importance_type='gain',\n", " interaction_constraints='', max_delta_step=0, max_depth=6,\n", " min_child_weight=1, missing=nan, monotone_constraints='()',\n", " n_estimators=100, n_jobs=0, num_parallel_tree=100,\n", " objective='binary:logistic', random_state=0, reg_alpha=0,\n", " scale_pos_weight=1, tree_method='exact', validate_parameters=1,\n", " verbosity=None)" ] }, "execution_count": 23, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf = xgboost.XGBRFClassifier()\n", "clf.fit(X_train, y_train)" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.801829268292683" ] }, "execution_count": 24, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf.score(X_test, y_test)" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.7790164452877926" ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "roc_auc_score(\n", " y_test, clf.predict(X_test)\n", ")" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[('sex_female', 0.45613697),\n", " ('sex_male', 0.4237464),\n", " ('pclass', 0.043916065),\n", " ('sibsp', 0.02123294),\n", " ('embarked_C', 0.011571336),\n", " ('fare', 0.011395166),\n", " ('parch', 0.008786586),\n", " ('embarked_S', 0.008488986),\n", " ('age', 0.0080112945),\n", " ('embarked_Q', 0.0067142267),\n", " ('embarked_', 0.0)]" ] }, "execution_count": 26, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sorted(zip(X_train.columns, clf.feature_importances_), key=lambda x: -x[1])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Optimize model" ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [], "source": [ "clf_ = xgboost.XGBRFClassifier()" ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [], "source": [ "params = {\n", " 'min_child_weight': [1, 5, 10],\n", " 'gamma': [0, 0.5, 1, 1.5, 2, 5],\n", " 'colsample_bytree': [0.6, 0.8, 1.0],\n", " 'max_depth': [4, 5, 6, 7],\n", "}" ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [], "source": [ "clf = model_selection.GridSearchCV(\n", " clf_, params, n_jobs=-1, \n", ").fit(X_train, y_train)" ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'colsample_bytree': 0.6, 'gamma': 1, 'max_depth': 7, 'min_child_weight': 1}" ] }, "execution_count": 30, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf.best_params_" ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.7957317073170732" ] }, "execution_count": 31, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf.score(X_test, y_test)" ] }, { "cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [], "source": [ "clf_best = xgboost.XGBRFClassifier(**clf.best_params_)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Confusion matrix" ] }, { "cell_type": "code", "execution_count": 33, "metadata": {}, "outputs": [], "source": [ "from sklearn.metrics import confusion_matrix" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "XGBRFClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n", " colsample_bytree=0.6, gamma=1, gpu_id=-1,\n", " importance_type='gain', interaction_constraints='',\n", " max_delta_step=0, max_depth=7, min_child_weight=1, missing=nan,\n", " monotone_constraints='()', n_estimators=100, n_jobs=0,\n", " num_parallel_tree=100, objective='binary:logistic',\n", " random_state=0, reg_alpha=0, scale_pos_weight=1,\n", " tree_method='exact', validate_parameters=1, verbosity=None)" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf_best.fit(X_train, y_train)" ] }, { "cell_type": "code", "execution_count": 35, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([[178, 26],\n", " [ 41, 83]])" ] }, "execution_count": 35, "metadata": {}, "output_type": "execute_result" } ], "source": [ "confusion_matrix(y_test, clf_best.predict(X_test))" ] }, { "cell_type": "code", "execution_count": 36, "metadata": {}, "outputs": [], "source": [ "import warnings\n", "warnings.simplefilter('ignore', FutureWarning)" ] }, { "cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "cm_viz = ConfusionMatrix(clf, classes=['died', 'survived'])\n", "cm_viz.fit(X_train, y_train)\n", "cm_viz.score(X_test, y_test)\n", "cm_viz.show();" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### ROC curve" ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.7709519291587603" ] }, "execution_count": 38, "metadata": {}, "output_type": "execute_result" } ], "source": [ "roc_auc_score(y_test, clf_best.predict(X_test))" ] }, { "cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "roc_viz = ROCAUC(clf)\n", "roc_viz.fit(X_train, y_train)\n", "roc_viz.score(X_test, y_test)\n", "roc_viz.show();" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Precision-recall curve" ] }, { "cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfYAAAFlCAYAAAAZGcpRAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAABC+UlEQVR4nO3deZyN5f/H8deZfbONLWRkGyqJIZGlbNn3LJEtSl+VilBKJLtWWlRKpcVWIYpCEvllyZKMUVos2cYy+5ztvn9/HE7GjDHGnFnOvJ+Px/f76Jz7Pvf9ORcP73Nd931fl8U0TRMRERHxCj55XYCIiIjkHAW7iIiIF1Gwi4iIeBEFu4iIiBdRsIuIiHgRBbuIiIgX8cvrAqTwOHLkCK1btyYyMtL9nmmaDBgwgHvuuSdHzvHaa69RqVIlunbtetl9unTpwoIFCyhatGiOnLNFixb4+/sTFBSExWLBZrPh4+PDmDFjaNasWY6c42I1atRgy5YtbNiwgTVr1vD222+n28dqtfLWW2+xYcMGTNPEMAw6derEAw88gMViyfGaMrNp0ybGjx9PeHg4n3zyCUFBQdk6TkpKCnPnzmXt2rVYLBaSk5OpV68eo0ePpkyZMun2P3HiBI899hgLFy5Mt23SpEmUKFGCRx99lMTERKZPn87u3buxWCz4+PjQr18/evbsma06r8acOXM4e/Yszz33XIbbf/nlF9544w1iY2NxOp2UL1+eJ598ksjISPr370+TJk0YNmxYms+8//77bN26lblz53q8fsmfFOySq4KCgli+fLn79YkTJ+jYsSO1atWiZs2a13z8xx577Ir7XHz+nPLiiy9yyy23uF+vXr2acePGsWnTphw/15WYpsnw4cOpXLkyixYtIjAwkLNnzzJs2DCSk5N5/PHHc7WeVatW0bNnT4YPH57tYzidToYOHUq1atVYsmQJISEhGIbBvHnzeOCBB1i2bFm6Hyxly5bNMNQv9dJLLxESEsKKFSuwWCycOHGC3r17U65cOZo0aZLtmq/Vtm3bGD16NK+//jq1atUCYMWKFfTv359vvvmGfv368corr6QL9sWLF/Pss8/mRcmSTyjYJU+VLVuWSpUq8ffff7Nv3z6WLl1KSkoKYWFhLFiwgCVLlvDZZ59hGAbFixdn/PjxVK1alaSkJCZPnswvv/yCr68vrVq14oknnuDpp5+mevXqDBkyhNmzZ/Pdd9/h7+9PiRIlmDZtGmXKlHH3eMPDw3njjTdYtWoVvr6+VK5cmfHjx1O6dGn69+9PnTp1+OWXXzh27Bj16tVjxowZ+Phc+eqVaZocOXKEYsWKud+72u/x999/M2nSJJKTkzl58iQ1a9bk1VdfJTAw8Irn37ZtG3/++SfvvPMOvr6+AJQoUYKZM2dy9OhRAPr370+/fv1o27Ztute1atWiZcuW7N+/n3vuuYft27e7RwUOHjzIoEGD2LBhA3///TdTpkzh3LlzOJ1O+vfvn27kZd68eaxbt47AwEASEhIYOXIk06dPZ8uWLfj6+lK7dm2efvppwsLCaNGiBbVr1yYmJoaRI0fSunVr93HWrl1LQkICEyZMcP8Z+Pj48OCDDwKQlJTEuXPn6NevH1WrVuXo0aNMnz6d+++/n507d5KYmMgzzzzD/v37KVOmDL6+vtSrVw+AU6dOUbJkSex2OwEBAZQtW5Y5c+ZQvHhxwPXjc9KkSRw7dgy73U6HDh146KGHANwjCFarlZSUFMaOHUvr1q2ZM2cOu3bt4uTJk9SoUYPp06cza9YsNmzYgK+vL3Xr1mXChAkA/Pnnn/Tv359Tp05RqlQpXn75ZcqUKcPs2bMZPny4O9QBOnfuTGBgIE6nk1atWjFlyhS2b99O/fr1Adi6dSumadK4ceMr/j0R76Vglzy1c+dODh06xK233sqWLVv4448/WL9+PWFhYWzdupVly5bxySefEBwczKZNm3j00Uf5+uuvmT17Nlarla+//hqn08n999/P1q1b3cc9duwYH374IVu2bCEgIID333+fPXv20KpVK/c+n3/+OT/++CNLly4lJCSEOXPm8NRTT/Hee+8BcOjQIRYsWEBycjLt2rVj69atNGzYMMPv8eSTTxIUFMS5c+cwTZMmTZq4h0Kz8z02bNhA165d6dKlC3a7ne7du7NhwwbatGlzxTbdu3cvtWvXdof6BTfccAM33HDDFT9vt9tp3rw5r732GomJibz99tucOnWK0qVL88UXX9C9e3dM02TEiBHMnDmTm2++mYSEBHr37k21atWoU6eO+1hDhw7ljz/+SPNj6+TJkyxfvhxfX1+eeeYZZs6cyaRJkwCoXr06r776arqatm/fzh133JHhD6sL4X7u3DmOHz/OSy+9RP369Tly5Ih7n9mzZxMUFMTq1as5e/Ys3bp1cwf7I488wmOPPUbDhg2pW7cuUVFRtG/fnooVKwIwevRoBg0aRIsWLbBarTzwwANERERw66238tNPP/Hxxx8TFBTEqlWrmD17tvsHydGjR1m5ciV+fn589NFH/PbbbyxfvpyAgABGjhzJ119/DcDhw4dZsmQJ4eHhDB8+nCVLlvDwww+zd+9ed/hf7OK/A71792bp0qXuYF+0aBF9+/bN9cstkr8o2CVXpaam0qVLF8A1vFqiRAlmzZpFuXLlANf147CwMAA2bNjAP//8Q58+fdyfj4uL49y5c/z00088/fTT+Pr64uvry8cffwzAl19+CbhGAmrWrEm3bt1o1qwZzZo1o1GjRmlq2bhxI927dyckJASAAQMGMHfuXGw2GwDNmzfHx8eHsLAwKlWqRFxc3GW/14Wh+MOHDzN48GCqVq3qDobsfI/bbruNzZs38+677/L3339z8uRJkpOTs9TGPj4+XOtM0ReCIiwsjDZt2rBixQoGDRrEihUr+PTTT/n77785dOgQ48aNc38mNTWVffv2pQn2S23cuJEnnngCf39/wDVS8PDDD6c776VM00wTVv/3f//HtGnTAFdbTpgwgerVq+Pn55fh+bds2cK4ceOwWCyEh4enGQ2oWbMmq1ev5rfffmPbtm1s3ryZuXPn8tprr9GwYUO2bdtGXFwcr732GgDJycns37+f9u3bM2PGDL766iv++ecfdu/eTVJSkvu4derUwc/P9U/sTz/9RJcuXdz3F1z48TJnzhwaN25MeHi4u5YzZ84Arj9HwzAu25YAvXr1okOHDiQmJuJwONi0aRMTJ07M9DPi/RTskqsuvcZ+qQshC2AYBl26dGH06NHu1ydPnqRYsWL4+fml+Yf+2LFjaW7K8vHx4eOPP+bXX39ly5YtTJ06ldtvvz3NtcdLw88wDBwOR5paL7BYLJimyWeffea+blurVi2mTJmS5hgVK1Zk5syZ9O/fn/r163Prrbdm63s8//zzOJ1O2rVrx1133cWxY8eyHNa33norH374IU6nM02vfc+ePSxYsIBZs2al+/52uz3NMS7+c+jZs6f70kG1atWoWLEiMTExFC1aNM2fZWxsLEWKFMm0tkuDyjCMNOe++LwXi4qKco+kADRs2NB97v79+2O1WgEICAhwh+mlLv6+F9rF4XDw/PPPM2rUKGrVqkWtWrUYPHgwb775JosWLaJBgwaYpsnChQsJDg4G4MyZMwQGBvLbb78xfPhwBg0aROPGjbntttt4/vnnM/wul9YUGxvrbouLt134ewauHwa7d+9Oc7MpwPPPP0/r1q254447KFOmDHfccQdff/01ycnJtGnT5op/BuL99Lib5FuNGzdm1apVnDx5EoDPPvuMgQMHAtCoUSO+/PJLDMPAZrMxYsQItm3b5v7s/v376dixI1WrVmXYsGEMGjSImJiYNMdv0qQJX3zxhbsnvGDBAm677TYCAgIuW9O9997L8uXLWb58ebpQvyAqKopu3brx/PPPYxhGtr7Hpk2bePjhh2nfvj0Wi4Xdu3fjdDqz1G5169alSpUqTJs2zR14sbGxTJ48meuvvx6A8PBw9u7dC7guOVzaNhe70AN+44033HeKV65cmcDAQHe4Hjt2jI4dO7qPeTlNmzZl4cKF2O12DMPgk08+ydL14LvvvpuQkBCmTJmSple8e/duDh8+nO6yQ0bnXbp0KYZhEBcXx7p16wBXqP7999+8+eab7h8YDoeDw4cPc9NNNxEWFkadOnWYP38+APHx8dx7772sW7eObdu2uX8INGjQgHXr1l32z6hRo0asXLkSm82GYRhMnDiRVatWZVrz//73P15//fU0bfrFF1+wZs2aNGHft29fvvrqK5YtW0a/fv0yPaYUDuqxS77VtGlTHnjgAe6//34sFgthYWG8/vrrWCwWHnnkEaZMmUKXLl1wOp20b9+eu+++m/Xr1wOuIc127drRo0cPQkJCCAoKSnen8D333MOxY8fo2bMnhmFQqVIlXnzxxRypfeTIkbRr145FixZx7733XvX3OHXqFA8//DDFihUjODiY2267jUOHDmX5/LNnz+aVV16he/fu+Pr6YhgGXbt2ZciQIYArNJ566il++OEHqlSpctkh8At69uzJm2++6b5HISAggDfffJMpU6Ywb948HA4Hjz32mPu69eX873//Y8aMGXTt2hWHw0Ht2rUZP378Fb+Pn58f8+bNY968edx3332YpklqairlypVjzJgxtGrVyn1jYEYeffRRJkyYQLt27QgPD08TjK+99hqzZs2iTZs2BAcHY5omrVq1cl8iePHFF3nhhRfo1KkTNpuNjh070rlzZ2JjY/n2229p3749/v7+NGrUiLi4OBITE9Odv0+fPhw9etR9f0KDBg3o378/b7311mVrrl+/PpMnT2bKlCkkJydjt9uJiIjgo48+olSpUu79br/9diZPnkyxYsWoUaPGFdtSvJ9Fy7aKiIh4Dw3Fi4iIeBEFu4iIiBfxaLDv3r2b/v37p3t//fr19OjRg969e7N48WJPliAiIlKoeOzmuXfffZcVK1a4HxG5wG63M23aNJYuXUpwcDD33nsvLVq0SHMziIiIiGSPx4I9IiKCOXPmMGbMmDTvHzx4kIiICPd0m/Xq1WPbtm20a9fusscyDIOkpCT8/f01o5KIiHg90zSx2+2EhoZmaSrri3ks2Nu0aZNmSscLEhMT00ygEBoamuHjIRdLSkriwIEDOV6jiIhIfhYZGXnVkw7l+nPsYWFhaSaYSEpKumLRF6afpHhZ8PWeR+9f+O5XigT6Ma/3HXldCuCaY/ziBSfEM9TOnqc29jy1sWfZbDYOHDjwX/5dhVxPyapVq/LPP/9w7tw5QkJC2L59u3vSjMtxD7/7+oHv1X/J/Oqs1cBpMbO0YlduyU+1eDO1s+epjT1Pbex52bn8nGvB/tVXX5GcnEzv3r156qmnGDJkCKZp0qNHD8qWLZtbZYiIiHg1jwb79ddf736crVOnTu73W7RoQYsWLTx5ahERkUJJE9SIiIh4EQW7iIiIF1Gwi4iIeBEFu4iIiBdRsIuIiHgRBbuIiIgXUbCLiIh4EQW7iIiIF1Gwi4iIeBEFu4iIiBfxnqXSvMDsjdGs/f2Y+7XDaeI0TQAsQPsbKzCi2Y15VJ2IiBQECvY8FJuUSrLNQef31gPwb1wKDsOgTFgQAHangeHKdc6mWFn7+zEFu4iIZErBnsdsToMUuxOAYkH+RF0fTp+6ldPt9+SK7bldmoiIFEAK9jzmMAymd4jK6zJERMRL6OY5ERERL6Ieex76qG+TvC5BRES8jHrsIiIiXkQ99kLmtY3RfBfzr/t1y8hyPHHnTXlYkYiI5CQFu5ebvTGaNTH/cv5xeI4npGB1OCkVGsSZZCur9/+rYBcR8SIKdi/0yg/7+C7GNdHNiYQUbE6D8JBAAIoE+tOy+nX0qVuZkcv1CJ2IiLdRsHuB2Ruj+fai4fWjcSnYnK5eeVigPw0iSmb4bLwn67l4Bj2AVtXLaXIdEZFcoGAvgC4NzuPxKSTbnZQ83ysvGuRPg4jrci3MM6on9fxwP8C5FJtmzRMRySUK9gIiNslK0vnpZ4/Hu4bXiwUFABDs78edVcvmaq/84pvwjiek4jAuX8/YlTty/PzpRgVM1xS8F7SMLMeTzW/O8fOKiOR3CvYCwjQh1eEk2eagaJA/t1UsyT233pBr578w3G9kcBNeaIBfrgz3z94YzdoDx9zntzkNSoW6RilMExznizuTbOW7A8cU7CJSKCnYCwgTE6cBMzrWy7VzZvUmvJyS0bV5h9N0//e/8cnYnQalQoPOjwpk/GOiIN8UmNFIxIUfLACtIq9j5F36wSIil6dgF7eF+0/z3C/r3I/GHYtPPt8r9txNeBcH2fH4FJyG6e6FA1gdBhdirVhQQK7fCOgJC/ef5vld692vTROc58P7REIKDsOk5EVt4LxoJGJNzDEFu4hkSsFeQHhi+tnYpFT8fX3oOG8dAH8ej8MnIPCim/ByPkgv7ZEei0vB6nRSJiwoTy4xeMql39M0XaMu4Gpn/8AgSoe52tm4KNgz+wFVkEciRCT3KNgLObvTICHVAUCwn4UWNcvneI/4VKKVQD+7+8Y/q8MgPMR1o11YoB93VczZG/9ik1IJ8PV1r3MPnn/cLqMfLHbDoEyY68kAh2G6wzvYz0KLGuWu+TvrsUIRyYiCXXi5S30AovdHc2PNnB/mdhoGyTaT+BQ7IZlcG89JdqeT+BQ7AKeTrXxnXvvjdhn1wh2G6078EwmpOC66ByE0wI/bLtPzvpZ2PpGQQtu317rPaTcMSp0/Z1yqHisUEQV7oZZbq8uZ5///pfM/IHLrnBfOl90h7Iyez7c6jAyH0IvkwkRADSJKsvXQaVLsTuDCfAUlPfpYoYgUPAp28bjcXp72as6X2V3oF25kKx588fP5eXfzXp+6lQv8jYMi4nkKdil0Lr0T326YlL7oeXjb+Ylu8mI63mthmHA0Ltk9VN+mZnkt8CNSCCnYxevFJqWSbHPQeZ7rZrqjca7H+EqHuR7jq1+xJL3r3JC3ReaAeteHu4fqzyRb+S7mWJaCPTduwtONfiK5R8EuhUKy3UFcqutmuoLWE8+qi4fqRy7fzqnE1DRPBlzMNME8P2HB8YRUbE4npc/P7X82m3P7p7+50LzosoZr2uGSF86RbNWNfiIeomCXQuPlXLx5L681iCjJL0fOkGJzZrjdaf73+F1ogB/NI7I2t/+l4W0YZppphu1Og1LnH/HD/G/bpTcXPrlCz+SLeIqCXbxebt+8lx9cy412F887AK7wvjD73/H4tOFtZhLeV3IiIZVO5y+PtKxejsfvVO9dJCco2EUkDcM0sToMkm2uiYucF/XKc+oyRv2Krkf34lPtnEm28s3+owp2kRyiYBeRNAzTxDBNjy44dOn9ACKScxTsIpJGYbx0IeJNfPK6ABEREck56rGLSL6iZ95Fro2CXUTy3MmEFNq9c9HiNk7DvSZ9XIpdz7yLXAUFu4jkqQuL2ySff+b+0sfm7l/4E8fjU9yP36n3LpI5BbuI5KkrPXNvmCY2p4HV7uRUkpVvD/ybpWDXkL4UVgp2EcnXDNPEcJpM7RDFyOXbMYz/tmU0E57TNImPiyfFJ8A1mc75If1zGtKXQkLBLiL52qWP351ISKHDu+uAy8+El+o0KBKs9eqlcFKwi0iBceF6fKLVNSve5WbCi94fzY011TOXwknBLiIFxrXMgS9SWGiCGhERES+iYBcREfEiCnYREREv4rFr7IZhMHHiRGJiYggICGDy5MlUqlTJvf39999n5cqVWCwWHnroIVq3bu2pUkREsiyj59/tToMLi9JbLBba1iyvx+Yk3/JYsK9duxabzcaiRYvYtWsX06dP56233gIgPj6ejz76iG+//ZaUlBS6du2qYBeRXJMuvE1wmq7kPpGQgsMwCQ8JdG92XliQHohLtel5eMnXPBbsO3bsoGnTpgDUqVOHvXv3urcFBwdTvnx5UlJSSElJwWKxZOmYRt/OEHsqzXuW+4dj6Xe/a/uYh2HHz+k/eGsUPi+/A4C5eAHm269leHzLNz9hCQjA/PN3zAfuzXifF17CcsedrvP1agOnT6ffqWsvfB4d49pn5kRYszL9PtdXxOfDL101rfsGc+r4jM/3yQos15XHjDuH2b1Vxvs88TSWjj1c53uoH/wek36nps3xmTjLdb55r2N+9kG6XW7w9YVvXe1n7tqOOeqhjM83+z0sN9/qOl+bhuBwpN9nwANYBg5z7fPM4/B/m9If6KZb8Jkz33W+Lxdivv5ixudb8QOW0FDMw39jDron430mzMDSrKXrfP06wfFj6Xfq0BWfkc+69nl5Mqxaln6f68rh88lXrpo2rsN8fmzG5/tgKZaKN2AmJWF2vjPjfR55Eku3Pq7zPToY9v0KQBW7DcM/wLVTwyb4THnVdb4P38b86N30B/Lzw2fN/7n2+W035oghGZ/vpblY6tR3na9jM0hJTr/PvYOwDH3Etc/E0fDj9+kPVL0GPnM/cZ1v5eeYr0zL+HxfrMVSrDjm8X8x+3XOeJ9xL2Bp2c51voHd4Mjh9Du16YjPmImufebMhGWL0+9TsiQ+i9e4avrpB8zxozI+37ufYalSHex2jJb102wblZSKxWJhwk99+aZWc+xOk0e+fJmKh/a7jvtfdhNfK4oq737oev+T9zHff9O97WyKFYvFwu6Xgrh130EAkvf+yu+9umZYU5V35lOkSTMA9jaqhzPuXLp9Sg+8n/JjnwHgnycf59zXX6XbJ6h6JDWWfwPAmS8/5/AzYzI8303f/4R/2bLYT5xgX/M7Mtyn4pSZhHdz/ZsR06Udqb8fSLdP8fadqPTiqwD8O2MKpz58P812w2Zjb+ky1Nrimh8gYdNG/nxwcIbnq754GSG1bgFg901VM9znusdGUXbYcAD+HDqQhJ/S/5sRdtvtVP3wUwBOzZ/Hv7My/rt5y65ofAICSInZz4FuHTLc54Y33qZYc9e/qfvubIj91Kl0+5Tq258Kz04E4NC40Zxd9kW6fQIr3UDNb1zzKpxduYJDY57I8Hw3fvcDARWux3H2LL81rp/hPtdPnEzJXq7s+Wv4UBj+eIb7XYnHgj0xMZGwsDD3a19fXxwOB35+rlOWK1eODh064HQ6GTZsmKfKEBFxM02TFLuTYH8/7qxaktrlS8CpwHT7hZcIy+DTIgWDxTQv/p2ac6ZNm8att95K+/btAWjWrBkbN24EYN26dXzwwQfMmzcPgCFDhjBmzBhq166d4bGsVit79+7lj3Op2A2PlCsiXm7CT0cBeP6OCtd0nDk7TxDk68PLd0XkRFkimapVqxaBgel/fGbGYz32qKgovv/+e9q3b8+uXbuIjIx0bytWrBhBQUEEBARgsVgoUqQI8fHxVzxmtapVwdffUyUXepqtK3eonT0vozZenENtXuSPZEIC/KhXr16OHK+g2rFjR6FvA0+60KHNDo8Fe+vWrdm8eTN9+vTBNE2mTp3K/PnziYiIoGXLlvz000/06tULHx8foqKiaNy4sadKERERKTQ8Fuw+Pj5MmjQpzXtVq/5308SIESMYMWKEp04vIiJSKGmueBGRq3Aq0Uqgn53O760HtMa75D+aeU5E5Co4TROrwyDZ5uDQ2STWxPyb1yWJpKEeu4jIVTBNE4dpMqNjPQZ8uolj8Sl0mrfu/FYLrSPVg5e8pWAXEbkKH/Vtkua11eEkPtU1SdPZFKtmpZM8p2AXEblGL3dxzST25IrteVyJiIJdRCTbLu29i+QHunlORETEi6jHLiLiAbM3RvPdgbSLEenGOskNCnYRkRxiAkfPJdPm7bWcTEjFaRqUCHbN8x2blMqC7SlplovVM/DiCQp2EZEcclvFkmw9dJpUu5OiQf40iChJn7qVARjw6SasTidJVtcd9KeTrSzY/me6Xv0F6t1LdinYRURySJ+6ld1BnhGnYTKzk2vhlIU7/2LrodPEp9rT7afH5uRaKNhFRHLBpXfQZ/Yj4OLH5mZvjE4zfA8awpfMKdhFRPKZ2CQrSTYHnd9bz/H4FJyGSXhooHvb2gPqzcvlKdhFRPIZwzRJtjuJT7UTEuDHbRX/u1Y/cvl2jsYl0/bttQD4WCy0rVleQS9uCnYRkXzINE1e6lw/3fsNIlw36KXYnYDrevyH2w6yer9rMZq7a5Tj8TtvytVaJX9RsIuI5DOZzWh36bX5CzfhJdkcnEm28m3MMQV7IadgFxEpwC4O+pHLNVe9aEpZERERr6JgFxER8SIKdhERES+ia+wiIl4iNimVBKsvneatd7+nqWkLH/XYRUS8iNXhev49PtXOP2cT081aJ95PPXYRES/zchfX8++DPtvM8fgUOr/n6sFrKtrCQcEuIuIlLn3+3QRsTgOr3cmpJCtrYv5VsBcCGooXEfFSpmlidxpM7RBFkUB/HIaZ1yVJLlCPXUTES13cg49NSsXf14fOF91Y10o31nkl9dhFRAoJu9MgLtVOXKqdv84k8t0B3VjnjdRjFxEpRC7cWKfpZ72Xgl1EpBDIbGEZ8S4aihcREfEiCnYREREvomAXERHxIgp2ERERL6JgFxER8SK6K15EpBCKTUol2eZwzyMPmkveWyjYRUQKKbthkGp3AhCbZOXjHX+6V4NTyBdcGooXESmk7E6DaR2imNYhCpvTIMXuJMHq4J8zSayJ+Tevy5NsUo9dRKQQSrcSnGliAi92qsfI5dsxtV5MgaVgFxGRdAvGJFp9df29gNJQvIiIpOMwTFLtTlLtTo6eS3Zfe5f8Tz12ERFJx2G4rr8DDPpsM8fjU9w9eMMwOXsujqI7zgLQtmYFHlNvPt9QsIuISBrpr7+D1WGQbHMA4DQh1WngZ3NyJtnK2gPHFOz5iIJdREQyZWLiMExmdKznfi96fzQ31rwx3fKvszdGpxm217X53KdgFxGRTF1pyddj8Sm0f2cdAP/GJ2N3GpQKDSI2KZV/ziSxKvooABag/Y0VFPQepmAXEZFsaxBRkq2HTpN0fpi+WFAADSJK0qduZRbu/Iuth067J8E5k6xJcHKDgl1ERLKtT93K9KlbOUvbBn66iRS7k0Srg9NJVr6JPqpg9wAFu4iI5AoT10Q4s85PgmNctG32xmi+2X/0shPjtKlZnifuvCk3yizwFOwiIpIrLp0Ex9/Xh07zXI/QHU9IIcXupGRIYLrPnUm28l3MMQV7FinYRUQkT9idBvGpdgBC/P24q2rZDIf1L73zXjKnYBcRkTzzcpf6eV2C11Gwi4hIrrvSI3SSfZorXkRExIt4rMduGAYTJ04kJiaGgIAAJk+eTKVKldzbf/jhB9544w1M0+Tmm29mwoQJWCwWT5UjIiIFlFabuzoe67GvXbsWm83GokWLGDVqFNOnT3dvS0xMZNasWcydO5clS5ZQoUIFzp4966lSRESkgLMbBik2Jyk2J0e02lymPNZj37FjB02bNgWgTp067N27171t586dREZGMmPGDA4fPkzPnj0JDw/3VCkiIlLAOQ2T6R1dq809uUJ3yWfGY8GemJhIWFiY+7Wvry8OhwM/Pz/Onj3Lzz//zLJlywgJCaFfv37UqVOHypUznr3ogj8OHsRuXGb2AskR0fuj87qEQkHt7HlqY8/LrTZ+OqpkmvMlJCZipPqwY8eOXDl/QeOxYA8LCyMpKcn92jAM/PxcpytevDi33HILpUuXBqB+/fpER0dfMdirVa0Kvv6eKrnQu7Bak3iW2tnz1Mael5dtXORAEkWC/KlXr96Vdy6grFZrmpHuq+Gxa+xRUVFs3LgRgF27dhEZGenedvPNN3PgwAHOnDmDw+Fg9+7dVKtWzVOliIiIFzGBf+NcK8q1f2cdL33/W16XlK94rMfeunVrNm/eTJ8+fTBNk6lTpzJ//nwiIiJo2bIlo0aNYujQoQC0bds2TfCLiIhczm0V/1tR7kyylW8PHGNU85vzuqx8w2Kal5tyP/9wD0mUrKCheA/S8GXuUDt7ntrY8/JLGw/4dBP+vj5cXzzE/d7dkeUL/KNwF3KvVq1aBAamnz8/M5p5TkRECjS70yAh1bUefGxSKgu2/5nmcbjC9sy7Zp4TEZEC7+Uu9d3zzqfYncSn2IlPsXP4bFKhe+ZdPXYRESmwMppz3sTkpfMhP3Zl4XskTsEuIiJeQ4vLaCheRETEq2Spx3706FE+/vhj4uLiuPgm+mnTpnmsMBEREbl6WQr2xx9/nPr161O/fn2twCYiIpKPZSnYHQ4HY8eO9XQtIiIiOepUopVAP7t7ydfC8OhbloK9Xr16rF+/niZNmhAQEODpmkRERHKE0zRJsTtJtDo4lZjKR3EHvf4Z9ywF++rVq/n444/TvGexWIiO1upJIiKSf5mmiQnM6lSPAZ9uwuowSLS6JrM5k2xl7e/HCmewb9q0ydN1iIiI5LhLH38zTJNZnVyrwnnruu5ZCvaUlBRef/11tmzZgtPppGHDhjz22GOEhIRc+cMiIiL5QGF5xj1Lz7FPmjSJlJQUpk6dyowZM7Db7UyYMMHTtYmIiMhVylKP/bfffmPFihXu18899xzt27f3WFEiIiKSPVnqsZumSXx8vPt1fHw8vr6+HitKREREsidLPfZBgwZxzz330KJFC0zT5Pvvv+fBBx/0dG0iIiJylbIU7D169OCWW25h27ZtGIbBnDlzqFGjhqdrExERkauU6VD8999/D8CyZcvYt28foaGhFClShOjoaJYtW5Yb9YmIiMhVyLTH/uuvv9K8eXN+/vnnDLd37drVEzWJiIhINmUa7CNGjADSruKWkJDA8ePHqV69umcrExERkauWpbvilyxZwtNPP82ZM2fo0KEDI0aM4JVXXvF0bSIiInKVshTsn332GWPHjmXlypW0bNmSr776ih9//NHTtYmIiMhVytJd8QDFixfnhx9+YMCAAfj5+WG1Wj1Zl4iIiMedTEj1uiVds9Rjr1atGsOGDePIkSM0atSIxx57jFq1anm6NhEREY+5rWIpwgL9SLY5OHQ2iW/2H83rknJElnrsU6dOZefOnVSvXp2AgAC6dOnCnXfe6enaREREPKZ33RvoXfcGAEYu345h5m09OSXTYF+0aBG9e/dm7ty5AGkee9u3bx+PPPKIZ6sTERGRq5LpULxpesnPFxERkUIi0x57nz59AHjooYf44YcfaNmyJWfOnGH9+vX06NEjVwoUERGRrMvSzXPjx4/n22+/db/++eeftR67iIhIPpSlm+f27t3LV199BUB4eDizZs2iU6dOHi1MRERErl6WeuyGYXDy5En369OnT+Pjk6WPioiISC7KUo/9oYceolu3btSrVw/TNNmzZw/PPPOMp2sTERHJFbFJqQT4+ronq4GCO2FNloK9U6dONGjQgF27duHn58f48eMpU6aMp2sTERHJNXbDICHVDsDpZCvfHThWIIM9S+PpNpuNL7/8knXr1tGgQQMWL16MzWbzdG0iIiK5xjRNXuxcnxc716dEcGBel5NtWQr2SZMmkZyczL59+/Dz8+PQoUMaihcREa/xUd8mfNS3SV6XkSOyFOy//fYbI0eOxM/Pj+DgYGbMmEF0dLSnaxMREZGrlKVgt1gs2Gw2LBYLAGfPnnX/t4iIiOQfWbp5bsCAAQwePJhTp04xZcoU1q5dy8MPP+zp2kREROQqZSnYmzVrRq1atfj5559xOp289dZb1KxZ09O1iYiIyFXKUrD369ePb775hmrVqnm6HhEREbkGWQr2mjVrsmzZMmrXrk1QUJD7/fLly3usMBEREbl6WQr23bt3s2fPnjTLuFosFtatW+exwkREROTqZRrsJ06c4IUXXiAkJISoqCiefPJJihYtmlu1iYiIyFXK9HG3cePGUaVKFcaMGYPdbmfatGm5VZeIiIhkwxV77O+99x4AjRo1omvXrrlRk4iIiGRTpsHu7++f5r8vfi0iIuLNTiamFsjV3rJ089wFmm1OREQKgwYRJdl59AzJNgcAsUlWPth2kG/2H+VEQiq+Fgvliga7928VmX9CP9Ng//3332nZsqX79YkTJ2jZsiWmaequeBER8Vp96lamT93K7tcLd/7F1kOnSbY5sTqcAMSdX+L1bIqVtb/nnyVeMw32NWvW5FYdIiIi+dbFQT/g000AvNylPgBPrtieZ3VlJNNgr1ChQm7VISIiUiBcuryrCRw9l0ybt9cC0Kr6dYxuUSsPKnPJ0upuIiIikrHbKpYkLNCfVLuTf+OSWfv78Tyt56punhMREZG0Lh6mH7k874flPRbshmEwceJEYmJiCAgIYPLkyVSqVCndPg8++CAtW7bk3nvv9VQpIiIiueZkQiod3v3v5nKn8d907D4WC21rlvfojXYeG4pfu3YtNpuNRYsWMWrUKKZPn55un1dffZX4+HhPlSAiIpKrGkSUpGiQP4lWh/t/KXan+3//xiez9vdjHq3BYz32HTt20LRpUwDq1KnD3r1702xfvXo1FovFvY+IiEhBd+ljcpcau3KHx2vwWLAnJiYSFhbmfu3r64vD4cDPz48DBw6wcuVKZs+ezRtvvJHlY/5x8CD2i4Y0JOdF74/O6xIKBbWz56mNPU9tfPUSEhKw+/qwY4fnAt5jwR4WFkZSUpL7tWEY+Pm5Trds2TJOnDjBwIEDOXr0KP7+/lSoUIFmzZplesxqVauCr6a19ZTo/dHcWDN/TLDgzdTOnqc29jy1cfYU+SOZkAA/6tWrl+l+Vqs13Uh3Vnks2KOiovj+++9p3749u3btIjIy0r1tzJgx7v+eM2cOpUqVumKoi4iIyJV5LNhbt27N5s2b6dOnD6ZpMnXqVObPn09ERESaaWpFREQKC8OEf+NSaHd+Mpu7a5bniTtvytFzeCzYfXx8mDRpUpr3qlatmm6/Rx991FMliIiI5Cv1rg93zTlvdxKblMrH2//k+z/+m9AmJ1aQ08xzIiIiuaRP3cq83KW+e555m9Mgyeogyerg8NmkHHkUTjPPiYiI5BHDNJnZyXUj3ZMrtnMsPoUO766jeIAPI2sVzdYxFewiIiJ54NLFZOpXLMnWQ6dJtDrwv4YBdQW7iIhIPpBmchunHU4fzdZxdI1dRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iIJdRETEiyjYRUREvIiCXURExIso2EVERLyIgl1ERMSLKNhFRES8iJ+nDmwYBhMnTiQmJoaAgAAmT55MpUqV3Ns/+OADVq1aBcCdd97JI4884qlSRERECg2P9djXrl2LzWZj0aJFjBo1iunTp7u3HT58mBUrVrBw4UIWL17Mpk2b2L9/v6dKERERKTQ81mPfsWMHTZs2BaBOnTrs3bvXve26665j3rx5+Pr6AuBwOAgMDPRUKSIiIoWGx4I9MTGRsLAw92tfX18cDgd+fn74+/sTHh6OaZrMnDmTm266icqVK1/xmH8cPIjdMD1VsgDR+6PzuoRCQe3seWpjz1Mbe46/j4VqxYOy9VmPBXtYWBhJSUnu14Zh4Of33+msVivjxo0jNDSUCRMmZOmY1apWBV//HK9VXKL3R3NjzRvzugyvp3b2PLWx56mNPcxph9NHs/VRj11jj4qKYuPGjQDs2rWLyMhI9zbTNBk+fDg1atRg0qRJ7iF5ERERuTYe67G3bt2azZs306dPH0zTZOrUqcyfP5+IiAgMw2Dr1q3YbDZ+/PFHAEaOHEndunU9VY6IiEih4LFg9/HxYdKkSWneq1q1qvu/f/31V0+dWkREpNDSBDUiIiJeRMEuIiLiRRTsIiIiXkTBLiIi4kUU7CIiIl5EwS4iIuJFFOwiIiJeRMEuIiLiRRTsIiIiXkTBLiIi4kUU7CIiIl5EwS4iIuJFFOwiIiJeRMEuIiLiRRTsIiIiXkTBLiIi4kUU7CIiIl5EwS4iIuJF/PK6gGtmOME087oKr+DnYwGnI6/LSMtiAR/fvK5CRKTAKNDBHmTYKF00lAD/Av018o3ri9xIYGBQXpeRhs3u4FR8Eqk+AXldiohIgVBwE9FwUrpoKKGhIXldiddwGk78A/JXgF6o53B8qnruIiJZUHCvsZumeuqFRIC/ny63iIhkUcENdhEREUlHXd5rsGLZl/z911+MeGJkmvefGv0kL0ydir+/Z4a1/z16lN49ulHzxpuwWMBqtVG/QQMefexx5r75Bqu/XkXp0mUAiDt3jrvbtWPog8M8UouIiOQvCnYPmD7rRY+fo3LVqrw7/wMADMNgcP/7OBATA8B9AwZyT6/eANhsNu7p0pnuPe4hvGRJj9clIiJ5y2uC/c3NMWz443iOHvOuatcxvHGNTPfZs3s3w4beT1JiIsOGP0zTZnfSoU1rvlixkqmTnsc/IIB//z1K7KlYnp88hRtvuomFn37C+nVrSU1JoXjxErz02mt8s2oVy7/8EtM0GDrsIZZ9/jkzX34FgMH9+zHzpVcoXaZMhjVYrVbsdhtBwenvaI87dw6Hw05gUNptv+7Zw4szpmOaBqXLlGXK9Bk8POxBnp3wPJWrVGHp4kXExsbSuUtXHn/kYYoVL07jpk1ZsWwZny9fgcViYfqUyTRo2JCKFSOYNX0qpgnFihdjwqTJFClSJJutLiIi18Jrgj2vBAcHM/vNtzh75gwD+t1L4yZN02wvV748z06YyBdLl/DF0iU8/ex44uLimPvue/j4+DB82AP8tncvAEWLFuWVOa9jmiYvTp9OfFwcp06donjxEulC/a+DB3lg8CDXY94+vtzb7z4iIioB8PFHH7L66685cfw4pcuWYfzzkwgNDU3z+SmTJjJ15iyqVKnKsi8+568//7zsd4w9Hcsnixfj7x9A9L59/LJjB7fUrs32bVt5cuxT3D+wPxMnvUCVqtVY9sXnfDj/fR4Z8VgOtK6IiFwtrwn24Y1rXLF37Ql1oqKwWCyElyxJWFgYcefOpdles+aNAJS97jp279yJj48P/v7+PD1mNCEhIZw8cQKHwzUpTKXKNwBgsVho37Ejq7/5mqNHjtC1e/d05714KP5SF4bi9/32G0+PeZJKlW5It8/p2FiqVKkKQNfuPdJtNy+6C71ChQru+wW697iHlSuWczo2ljvvao6fnx9///kn0yZPBsDhsLt/YIiISO7TXfHXaN/53nZs7CmSk5MpXqJE2h0saV8eiIlhw/p1zHjxJcY8PQ7DMN0h6mP574+jc9durP12Db/s2E7jps2yVdtNN9/M4CFDeWrMkxiGkWZb6TJlOPTPPwB88N481q9bS0BAILGxpwDYH73vv6/g819dDRo2JGZ/NMuXfUHXHq4fBJVuuIFJU6fy7vwPeGzkKJreeWe26hURkWvnNT32vJJqTeXBIYNJSU7mmecmYrFYMt2/YkQEQcHBDO7fD4BSpUtx6uSpdPuVKVuWkNBQbql9K35+2f9j6tq9B9+uXs2SRQvpfW9f9/vPPDeBic89i4/Fh1KlS9O3/wDAZNrkyVxX7jrKlCmb4fEsFgstW9/Nz/+3hYoVIwB4+tnneG7cOJxOB1gsTHj+hWzXKyIi18Zimvl/5g+r1crevXuhZAXw9Xe96XRQtVSRfDdTWk4a8fBwnhw7NteGtlNTUwgKCs6Vc10Nu83GwdgE8PWO36HR+6O58fwlGvEMtbHnqY09zGmH00epVasWgYGBV/VRDcXnQ6mpqfTt1ZPKVaroerWIiFwV7+gCeZmgoCA+Xbwkr8sQEZECSD12ERERL6JgFxER8SIKdhERES+iYBcREfEiCvYc8sH773F38zuxWq3ZPsaKZV/SvnVLHhg8iAfvH8TAfn35dvXqa6pr1OOXn9p1/rx32fvrnms6/sXefutN+t/bm0H39cvwuF+vXEnfXvdwX59eLFm0EHAtYDNl0vMM7NeXBwYP4tChf3KsHhGRwkh3xeeQr1eu5O627Vjzzdd07tot28dp276DexnYuLhz9O7endZt2lxx4pvLeenV1y67bfDQB7J1zIxE79vHju3b+OjThRw/fozRTzzOxwsXp9nn1ZdmsWTZckJCQujRpTNt2rZj27at2KxWPvzkU/bs3s0rs2bxypzXc6wuEZHCxquC/fdbIzN8v+SjIwkf+hAARx8aTPKWzen2Ca5/O9e/twCAsx++R+zLM6i++0CWzrt921YqVqzIPb168+zTY6l54028OGMa77z/AeCaaGb4I49y8uQJ5r7xOmFhYRQpWozqkZE8NPzhyx43IT6BwKBALBYLPbt1IaLSDfj7+/PMcxOYNGG8e1760U+No3pkJMu++JwlixZhGE6a3dWc/z38CK3vasZ3GzayeOFnrFyxHIvFh5tr1WLM0+OY8Mw47m7Xjga3N2Ti+Gc5fOgQYNJvwEDatG3HA4MHEVmzBgd//4OkpERmvPQKDoedFyZMSFNn2/btsdlsNGp0BxaLhXLlyuN0Ojl75gwlwsPd+1WLjCQxIRFfXz8wXbPY7frlF+5o0gSA2rfeyr59v2WpzUVEJGNeFex55cvPP6drjx7cULkyAQEBWK1WrFYb//77L/7+/pw7d5bqkZGMfuJxPvj4E0qWKsUzY8dkeKzVX6/i1z17sPhYCAoK5oWp0wFITk7mgWEPUfPGG3nt5ZdocHtDevbuw6F//mHi+Gd48ZXXmP/ePBZ9/iWBgYHMefUVkpOT3MddsexLnn52PDfXuoUlixa6F54B+HzJYkqUKMH4iRNxOg369rqHBrc3BKBWrVsYPfZpXp/9Gmu+XsXgoQ9kuPjMvLfnUqx4cffrkJBQEhIT0gZ7ter0692T4OAQWrRqRZGiRUlKSiIs7L8lXn19fHA4HNc0ja6ISGHmVf96ZqWHXWHu/CvuU2LgEEoMHJKlc8bHxbH5x42cPXOGhZ9+QmJCIos++5Su3buzasVy/AMC6NylG2fPniE0LJSSpUoBULdePWJjY9Md7+Kh+EtVuuEGAP74/Xe2bf3Zff09Li6OI0eOUK1adYLOr7t+6TEmvjCFBR/M5+jRF6l9661pVm/7688/ub1RIwBCQ0OpUqUqRw4fBqDGja4pI6+77jpiY2M5dOifDHvsoWFhJCX990MiOTmJIkWKul8fiIlh08aNfLX6W0JCQnj26bF8t2YNoaGhaT5nGKZCXUTkGuhf0Gv09cqVdOnegydGPQlASkoKndq24ZHHHmPMyCewWHx48513CQ4OJjkp2T08/evu3ZSrUOGqzuVzfpW1GypXpn3HjrTr0JEzp0/z5RefU7FiRf766y9sNhsBAQGMHvk4o8c+7f7sl58vZdxzEwgMDGT4sAfYs2uXe1vlKlXYuWMHdzRuTFJSEn/8/jsVztdmuWR5uoiIShn22Pf99huvvfISAwYN5sSJ4xiGQYmLVroLK1KEwKBAgoIC8fX1JTy8JPHxcdSpW5eNGzZwd9u27Nm9m2rVq19Vm4iISFoK9mv05RdL3cPlAMHBwbRs3YpVX31F9cgaOJ1OQkNDARg77hkeHf4QYWFFMEyDipWyNw/8kAeHMem58XyxdClJSYkM+99wSoSHM+j++xk6eCAWLDS76y7KlP1vhbZq1aszZGB/QkJDKVOmLLVq12bFsi8B6NGzJy9MnMBDQ4dgs9l58H//I7xkyauq6aabb6ZuVD0G3dcXwzB46plnAfhm1UqSk5Pp0bMXPXr24v4B/fH39+f6ihXp3LUrvr5+/N+WLQy6rx+maTLxhcnZahMREXHR6m656P1573LfgIEEBATwzFNjaXTHHXTs3CWvy3LT6m65Q6tieZ7a2PPUxh52Dau7ece/lAVESEgIA/vdS1BQEOXLV+Dutu3yuiQREfEyCvZc1KdvP/r07ZfXZYiIiBfTzHMiIiJepOAGu8WCze648n5S4NnsDsjmzHsiIoVNwR2K9/HlVLzr+ecA/4L7NfITh92O3cc3r8tIw2Z3uP6cfQrOTZIiInmpQCdiqk8Ah+NTIf/f2F8g/H7wD6pXrZbXZaRlsSjURUSuQoEOdgDyWQ+zIHMYptc8UiYiUlh57Bq7YRg899xz9O7dm/79+/PPP2mX41y8eDHdu3enV69efP/9954qQ0REpFDxWPds7dq12Gw2Fi1axK5du5g+fTpvvfUWAKdOnWLBggV8/vnnWK1W+vbtS+PGjQkoQJPNiIiI5EceC/YdO3bQtGlTAOrUqeOaOe68PXv2ULduXQICAggICCAiIoL9+/dTu3btDI91YXK8AAv4+Oh6uqeE+fkQpPb1OLWz56mNPU9t7FmGATYgO5PDeizYExMTCQsLc7/29fV1L8eZmJhIkSL/LdUZGhpKYmLiZY9lt9sBsMUe9VS5AlQsGkjq8X+uvKNcE7Wz56mNPU9tnDvsdrt71c6s8liwh12yjKdhGO7lOC/dlpSUlCboLxUaGkpkZCT+/v5Y9DyziIh4OdM0sdvt7kXErobHgj0qKorvv/+e9u3bs2vXLiIjI93bateuzauvvorVasVms3Hw4ME02y/l4+OTafCLiIh4m6vtqV/gsdXdDMNg4sSJHDhwANM0mTp1Khs3biQiIoKWLVuyePFiFi1ahGmaDBs2jDZt2niiDBERkUKlQCzbKiIiIllTcOeKFxERkXQU7CIiIl4k3wW7ZqzzvCu18QcffEDPnj3p2bMnr7/+eh5VWbBdqY0v7DN06FA+++yzPKiw4LtSG//www/06tWLnj17MnHixGw9DyxXbuf333+f7t2706NHD7777rs8qtI77N69m/79+6d7f/369fTo0YPevXuzePHiKx/IzGfWrFljjh071jRN09y5c6f50EMPubedPHnS7Nixo2m1Ws34+Hj3f8vVyayNDx06ZHbr1s10OBymYRhm7969zejo6LwqtcDKrI0veOmll8yePXuan376aW6X5xUya+OEhASzQ4cO5unTp03TNM133nnH/d9ydTJr57i4OPPOO+80rVaree7cOfOuu+7KqzILvHfeecfs2LGj2bNnzzTv22w2s1WrVua5c+dMq9Vqdu/e3Tx16lSmx8p3PfaszlhXpEgR94x1cnUya+PrrruOefPm4evri8ViweFwEBgYmFelFliZtTHA6tWrsVgs7n3k6mXWxjt37iQyMpIZM2bQt29fSpUqRXh4eF6VWqBl1s7BwcGUL1+elJQUUlJSNM/INYiIiGDOnDnp3j948CAREREUK1aMgIAA6tWrx7Zt2zI9Vr5byisnZ6yTjGXWxv7+/oSHh2OaJjNnzuSmm26icuXKeVhtwZRZGx84cICVK1cye/Zs3njjjTyssmDLrI3Pnj3Lzz//zLJlywgJCaFfv37UqVNHf5ezIbN2BihXrhwdOnTA6XQybNiwvCqzwGvTpg1HjhxJ9352ci/fBXtOzlgnGcusjQGsVivjxo0jNDSUCRMm5EWJBV5mbbxs2TJOnDjBwIEDOXr0KP7+/lSoUIFmzZrlVbkFUmZtXLx4cW655RZKly4NQP369YmOjlawZ0Nm7bxx40ZOnjzJunXrABgyZAhRUVGXXfdDrl52ci/fDcVHRUWxceNGgAxnrNuxYwdWq5WEhIQrzlgnGcusjU3TZPjw4dSoUYNJkybh66v17rMjszYeM2YMS5YsYcGCBXTr1o1BgwYp1LMhsza++eabOXDgAGfOnMHhcLB7926qVauWV6UWaJm1c7FixQgKCiIgIIDAwECKFClCfHx8XpXqlapWrco///zDuXPnsNlsbN++nbp162b6mXzXY2/dujWbN2+mT58+7hnr5s+f756xrn///vTt2xfTNHniiSd0/TcbMmtjwzDYunUrNpuNH3/8EYCRI0de8S+SpHWlv8dy7a7UxqNGjWLo0KEAtG3bVp2AbLpSO//000/06tULHx8foqKiaNy4cV6X7BW++uorkpOT6d27N0899RRDhgzBNE169OhB2bJlM/2sZp4TERHxIvluKF5ERESyT8EuIiLiRRTsIiIiXkTBLiIi4kUU7CIiIl4k3z3uJiKeceTIEdq2bUvVqlUB10QjSUlJdO3alREjRuTIOS5Mifnoo49So0YNYmJicuS4IpJ1CnaRQqRMmTIsX77c/frEiRO0adOGDh06uANfRAo2BbtIIXbq1ClM0yQ0NJR33nmHb775BqfTSZMmTRg9ejQWi4UPPviAzz77DF9fX5o3b87o0aM5cOAAL7zwAsnJyZw5c4bBgwczYMCAvP46IoKCXaRQOXnyJF26dMFqtXL27FluueUWXn/9dQ4cOMDevXtZunQpFouF0aNHs2LFCipXrsynn37K559/TnBwMEOHDmXv3r0sX76c4cOH06hRIw4fPkznzp0V7CL5hIJdpBC5MBRvGAbTp08nJiaGhg0b8vLLL7Nnzx66d+8OQGpqKuXLlyc2NpbmzZu7F5344IMPALjxxhv58ccfefvtt4mJiSE5OTmvvpKIXELBLlII+fj4MGbMGLp27cr777+P0+lk4MCBDB48GID4+Hh8fX1ZunRpms+dOHGC4OBgnnnmGYoWLUrz5s1p3749q1atyouvISIZ0ONuIoWUn58fY8aMYe7cudx0000sX76cpKQkHA4HDz/8MGvWrKF+/fps3LjR/f6oUaPYu3cvmzdvZsSIEbRq1Ypt27YB4HQ68/gbiQioxy5SqDVr1ow6deqwbds27r77bnr16oXT6aRp06Z069YNi8XCfffdR58+fTAMg9atW3PHHXfw6KOP0rdvX4oWLUrlypWpUKECR44cyeuvIyJodTcRERGvoqF4ERERL6JgFxER8SIKdhERES+iYBcREfEiCnYREREvomAXERHxIgp2ERERL6JgFxER8SL/D63gHp4JIK6PAAAAAElFTkSuQmCC\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "prc_viz = PRCurve(clf)\n", "prc_viz.fit(X_train, y_train)\n", "prc_viz.score(X_test, y_test)\n", "prc_viz.show();" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Learning curve" ] }, { "cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "lc_viz = LearningCurve(clf_best)\n", "lc_viz.fit(X_train, y_train)\n", "lc_viz.score(X_test, y_test)\n", "lc_viz.show();" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Model persistence (and deploymnet)" ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [], "source": [ "import joblib" ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "['clf_best.pickle']" ] }, "execution_count": 43, "metadata": {}, "output_type": "execute_result" } ], "source": [ "joblib.dump(clf_best, 'clf_best.pickle')" ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [], "source": [ "clf = joblib.load('clf_best.pickle')" ] }, { "cell_type": "code", "execution_count": 45, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0.7957317073170732" ] }, "execution_count": 45, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf.score(X_test, y_test)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.5" } }, "nbformat": 4, "nbformat_minor": 4 }