From d4b550fd78a243ba00925548d8a0276af86f3f92 Mon Sep 17 00:00:00 2001 From: rasbt Date: Sat, 7 Dec 2019 13:17:51 -0600 Subject: [PATCH] finish docs --- docs/sources/CHANGELOG.md | 2 +- .../classifier/StackingCVClassifier.ipynb | 326 +++++++++++++++--- .../classifier/StackingClassifier.ipynb | 320 ++++++++++++++--- 3 files changed, 542 insertions(+), 106 deletions(-) diff --git a/docs/sources/CHANGELOG.md b/docs/sources/CHANGELOG.md index fa9b94889..64a6ef8a3 100755 --- a/docs/sources/CHANGELOG.md +++ b/docs/sources/CHANGELOG.md @@ -19,7 +19,7 @@ The CHANGELOG for the current development version is available at - The `SequentialFeatureSelector` now supports using pre-specified feature sets via the `fixed_features` parameter. ([#578](https://github.com/rasbt/mlxtend/pull/578)) - Adds a new `accuracy_score` function to `mlxtend.evaluate` for computing basic classifcation accuracy, per-class accuracy, and average per-class accuracy. ([#624](https://github.com/rasbt/mlxtend/pull/624) via [Deepan Das](https://github.com/deepandas11)) -- `StackingClassifier` and `StackingCVClassifier`now supports `decision_function` method, which serves as a preferred choice over `predict_proba` in calculating roc_auc and average_precision scores when the meta estimator is a linear model or support vector classifier. +- `StackingClassifier` and `StackingCVClassifier`now have a `decision_function` method, which serves as a preferred choice over `predict_proba` in calculating roc_auc and average_precision scores when the meta estimator is a linear model or support vector classifier. ([#634](https://github.com/rasbt/mlxtend/pull/634) via [Qiang Gu](https://github.com/qiagu)) ##### Changes diff --git a/docs/sources/user_guide/classifier/StackingCVClassifier.ipynb b/docs/sources/user_guide/classifier/StackingCVClassifier.ipynb index bbdd12801..50d9f9072 100644 --- a/docs/sources/user_guide/classifier/StackingCVClassifier.ipynb +++ b/docs/sources/user_guide/classifier/StackingCVClassifier.ipynb @@ -108,9 +108,9 @@ "3-fold cross validation:\n", "\n", "Accuracy: 0.91 (+/- 0.01) [KNN]\n", - "Accuracy: 0.90 (+/- 0.03) [Random Forest]\n", - "Accuracy: 0.92 (+/- 0.03) [Naive Bayes]\n", - "Accuracy: 0.91 (+/- 0.01) [StackingClassifier]\n" + "Accuracy: 0.95 (+/- 0.01) [Random Forest]\n", + "Accuracy: 0.91 (+/- 0.02) [Naive Bayes]\n", + "Accuracy: 0.93 (+/- 0.02) [StackingClassifier]\n" ] } ], @@ -160,7 +160,7 @@ "outputs": [ { "data": { - "image/png": "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\n", + "image/png": "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\n", "text/plain": [ "
" ] @@ -228,9 +228,9 @@ "3-fold cross validation:\n", "\n", "Accuracy: 0.91 (+/- 0.01) [KNN]\n", - "Accuracy: 0.93 (+/- 0.05) [Random Forest]\n", - "Accuracy: 0.92 (+/- 0.03) [Naive Bayes]\n", - "Accuracy: 0.95 (+/- 0.04) [StackingClassifier]\n" + "Accuracy: 0.95 (+/- 0.01) [Random Forest]\n", + "Accuracy: 0.91 (+/- 0.02) [Naive Bayes]\n", + "Accuracy: 0.95 (+/- 0.02) [StackingClassifier]\n" ] } ], @@ -282,16 +282,16 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.673 +/- 0.01 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.933 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.673 +/- 0.01 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.927 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Best parameters: {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "Accuracy: 0.94\n" + "0.947 +/- 0.03 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.933 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "Best parameters: {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "Accuracy: 0.95\n" ] } ], @@ -352,23 +352,23 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.673 +/- 0.01 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.947 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.673 +/- 0.01 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.933 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.673 +/- 0.01 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.933 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.673 +/- 0.01 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.960 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.927 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Best parameters: {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "Best parameters: {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", "Accuracy: 0.96\n" ] } @@ -447,19 +447,39 @@ "data": { "text/plain": [ "StackingCVClassifier(classifiers=[Pipeline(memory=None,\n", - " steps=[('columnselector', ColumnSelector(cols=(0, 2), drop_axis=False)), ('logisticregression', LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n", - " intercept_scaling=1, max_iter=100, multi_class='warn',\n", - " n_jobs=None,...nalty='l2', random_state=None, solver='warn',\n", - " tol=0.0001, verbose=0, warm_start=False))])],\n", - " cv=2, drop_last_proba=False,\n", - " meta_classifier=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n", - " intercept_scaling=1, max_iter=100, multi_class='warn',\n", - " n_jobs=None, penalty='l2', random_state=None, solver='warn',\n", - " tol=0.0001, verbose=0, warm_start=False),\n", - " n_jobs=None, pre_dispatch='2*n_jobs', random_state=42,\n", - " shuffle=True, store_train_meta_features=False, stratify=True,\n", - " use_clones=True, use_features_in_secondary=False,\n", - " use_probas=False, verbose=0)" + " steps=[('columnselector',\n", + " ColumnSelector(cols=(0, 2),\n", + " drop_axis=False)),\n", + " ('logisticregression',\n", + " LogisticRegression(C=1.0,\n", + " class_weight=None,\n", + " dual=False,\n", + " fit_intercept=True,\n", + " intercept_scaling=1,\n", + " l1_ratio=None,\n", + " max_iter=100,\n", + " multi_class='auto',\n", + " n_jobs=None,\n", + " penalty='l2',\n", + " random_state=None,\n", + " solver='lbfgs',\n", + " tol=0.0...\n", + " fit_intercept=True,\n", + " intercept_scaling=1,\n", + " l1_ratio=None,\n", + " max_iter=100,\n", + " multi_class='auto',\n", + " n_jobs=None,\n", + " penalty='l2',\n", + " random_state=None,\n", + " solver='lbfgs',\n", + " tol=0.0001, verbose=0,\n", + " warm_start=False),\n", + " n_jobs=None, pre_dispatch='2*n_jobs', random_state=42,\n", + " shuffle=True, store_train_meta_features=False,\n", + " stratify=True, use_clones=True,\n", + " use_features_in_secondary=False, use_probas=False,\n", + " verbose=0)" ] }, "execution_count": 8, @@ -494,13 +514,185 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# API" + "## Example 5 -- ROC Curve with `decision_function`" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Like other scikit-learn classifiers, the `StackingCVClassifier` has an `decision_function` method that can be used for plotting ROC curves. Note that the `decision_function` expects and requires the meta-classifier to implement a `decision_function`." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, + "outputs": [], + "source": [ + "from sklearn import model_selection\n", + "from sklearn.linear_model import LogisticRegression\n", + "from sklearn.neighbors import KNeighborsClassifier\n", + "from sklearn.svm import SVC\n", + "from sklearn.ensemble import RandomForestClassifier\n", + "from mlxtend.classifier import StackingCVClassifier\n", + "from sklearn.metrics import roc_curve, auc\n", + "import numpy as np\n", + "from sklearn.model_selection import train_test_split\n", + "from sklearn import datasets\n", + "from sklearn.preprocessing import label_binarize\n", + "from sklearn.multiclass import OneVsRestClassifier\n", + "\n", + "\n", + "iris = datasets.load_iris()\n", + "X, y = iris.data[:, [0, 1]], iris.target\n", + "\n", + "\n", + "# Binarize the output\n", + "y = label_binarize(y, classes=[0, 1, 2])\n", + "n_classes = y.shape[1]\n", + "\n", + "\n", + "\n", + "RANDOM_SEED = 42\n", + "\n", + "\n", + "X_train, X_test, y_train, y_test = train_test_split(\n", + " X, y, test_size=0.33, random_state=RANDOM_SEED)\n", + "\n", + "clf1 = LogisticRegression()\n", + "clf2 = RandomForestClassifier(random_state=RANDOM_SEED)\n", + "clf3 = SVC(random_state=RANDOM_SEED)\n", + "lr = LogisticRegression()\n", + "\n", + "\n", + "sclf = StackingCVClassifier(classifiers=[clf1, clf2, clf3],\n", + " meta_classifier=lr)\n", + "\n", + "\n", + "# Learn to predict each class against the other\n", + "classifier = OneVsRestClassifier(sclf)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "**Using `predict_proba()`**" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "y_score = classifier.fit(X_train, y_train).predict_proba(X_test)\n", + "\n", + "# Compute ROC curve and ROC area for each class\n", + "fpr = dict()\n", + "tpr = dict()\n", + "roc_auc = dict()\n", + "for i in range(n_classes):\n", + " fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_score[:, i])\n", + " roc_auc[i] = auc(fpr[i], tpr[i])\n", + "\n", + "# Compute micro-average ROC curve and ROC area\n", + "fpr[\"micro\"], tpr[\"micro\"], _ = roc_curve(y_test.ravel(), y_score.ravel())\n", + "roc_auc[\"micro\"] = auc(fpr[\"micro\"], tpr[\"micro\"])\n", + "\n", + "plt.figure()\n", + "lw = 2\n", + "plt.plot(fpr[2], tpr[2], color='darkorange',\n", + " lw=lw, label='ROC curve (area = %0.2f)' % roc_auc[2])\n", + "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n", + "plt.xlim([0.0, 1.0])\n", + "plt.ylim([0.0, 1.05])\n", + "plt.xlabel('False Positive Rate')\n", + "plt.ylabel('True Positive Rate')\n", + "plt.title('Receiver operating characteristic example')\n", + "plt.legend(loc=\"lower right\")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "**Using `decision_function()`**" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "y_score = classifier.fit(X_train, y_train).decision_function(X_test)\n", + "\n", + "# Compute ROC curve and ROC area for each class\n", + "fpr = dict()\n", + "tpr = dict()\n", + "roc_auc = dict()\n", + "for i in range(n_classes):\n", + " fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_score[:, i])\n", + " roc_auc[i] = auc(fpr[i], tpr[i])\n", + "\n", + "# Compute micro-average ROC curve and ROC area\n", + "fpr[\"micro\"], tpr[\"micro\"], _ = roc_curve(y_test.ravel(), y_score.ravel())\n", + "roc_auc[\"micro\"] = auc(fpr[\"micro\"], tpr[\"micro\"])\n", + "\n", + "plt.figure()\n", + "lw = 2\n", + "plt.plot(fpr[2], tpr[2], color='darkorange',\n", + " lw=lw, label='ROC curve (area = %0.2f)' % roc_auc[2])\n", + "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n", + "plt.xlim([0.0, 1.0])\n", + "plt.ylim([0.0, 1.05])\n", + "plt.xlabel('False Positive Rate')\n", + "plt.ylabel('True Positive Rate')\n", + "plt.title('Receiver operating characteristic example')\n", + "plt.legend(loc=\"lower right\")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# API" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, "outputs": [ { "name": "stdout", @@ -656,6 +848,27 @@ "\n", "
\n", "\n", + "*decision_function(X)*\n", + "\n", + "Predict class confidence scores for X.\n", + "\n", + "**Parameters**\n", + "\n", + "- `X` : {array-like, sparse matrix}, shape = [n_samples, n_features]\n", + "\n", + " Training vectors, where n_samples is the number of samples and\n", + " n_features is the number of features.\n", + "\n", + "**Returns**\n", + "\n", + "- `scores` : shape=(n_samples,) if n_classes == 2 else (n_samples, n_classes).\n", + "\n", + " Confidence scores per (sample, class) combination. In the binary\n", + " case, confidence score for self.classes_[1] where >0 means this\n", + " class would be predicted.\n", + "\n", + "
\n", + "\n", "*fit(X, y, groups=None, sample_weight=None)*\n", "\n", "Fit ensemble classifers and the meta-classifier.\n", @@ -711,6 +924,11 @@ "\n", " Target values.\n", "\n", + "\n", + "- `**fit_params` : dict\n", + "\n", + " Additional fit parameters.\n", + "\n", "**Returns**\n", "\n", "- `X_new` : numpy array of shape [n_samples, n_features_new]\n", @@ -769,14 +987,14 @@ "\n", "**Parameters**\n", "\n", - "- `X` : numpy array, shape = [n_samples, n_features]\n", + "- `X` : {array-like, sparse matrix}, shape = [n_samples, n_features]\n", "\n", " Training vectors, where n_samples is the number of samples and\n", " n_features is the number of features.\n", "\n", "**Returns**\n", "\n", - "- `proba` : array-like, shape = [n_samples, n_classes]\n", + "- `proba` : array-like, shape = [n_samples, n_classes] or a list of n_outputs of such arrays if n_outputs > 1.\n", "\n", " Probability for each class per sample.\n", "\n", @@ -784,7 +1002,7 @@ "\n", "*score(X, y, sample_weight=None)*\n", "\n", - "Returns the mean accuracy on the given test data and labels.\n", + "Return the mean accuracy on the given test data and labels.\n", "\n", "In multi-label classification, this is the subset accuracy\n", "which is a harsh metric since you require for each sample that\n", @@ -792,17 +1010,17 @@ "\n", "**Parameters**\n", "\n", - "- `X` : array-like, shape = (n_samples, n_features)\n", + "- `X` : array-like of shape (n_samples, n_features)\n", "\n", " Test samples.\n", "\n", "\n", - "- `y` : array-like, shape = (n_samples) or (n_samples, n_outputs)\n", + "- `y` : array-like of shape (n_samples,) or (n_samples, n_outputs)\n", "\n", " True labels for X.\n", "\n", "\n", - "- `sample_weight` : array-like, shape = [n_samples], optional\n", + "- `sample_weight` : array-like of shape (n_samples,), default=None\n", "\n", " Sample weights.\n", "\n", @@ -875,5 +1093,5 @@ } }, "nbformat": 4, - "nbformat_minor": 1 + "nbformat_minor": 4 } diff --git a/docs/sources/user_guide/classifier/StackingClassifier.ipynb b/docs/sources/user_guide/classifier/StackingClassifier.ipynb index 0dd841141..a06fceee4 100644 --- a/docs/sources/user_guide/classifier/StackingClassifier.ipynb +++ b/docs/sources/user_guide/classifier/StackingClassifier.ipynb @@ -114,9 +114,9 @@ "3-fold cross validation:\n", "\n", "Accuracy: 0.91 (+/- 0.01) [KNN]\n", - "Accuracy: 0.93 (+/- 0.05) [Random Forest]\n", - "Accuracy: 0.92 (+/- 0.03) [Naive Bayes]\n", - "Accuracy: 0.95 (+/- 0.03) [StackingClassifier]\n" + "Accuracy: 0.95 (+/- 0.01) [Random Forest]\n", + "Accuracy: 0.91 (+/- 0.02) [Naive Bayes]\n", + "Accuracy: 0.95 (+/- 0.02) [StackingClassifier]\n" ] } ], @@ -160,7 +160,7 @@ "outputs": [ { "data": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAk0AAAHiCAYAAAD1WPj+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzs3Xd8W9X9//HX0bC8t5PYWc6ebDLKTBkhgQAp0BA2tDSMbxk/2lLooouWtpRCGYVQRhgBwoawEkagJJCEhISQQLYTO47jvaek8/tDtrEsyb6yJWt9no+HH42Pro8+8he/v+eee+65SmuNEEIIIYTomSnUBQghhBBCRAIZNAkhhBBCGCCDJiGEEEIIA2TQJIQQQghhgAyahBBCCCEMkEGTEEIIIYQBMmgSQggR0ZRSs5RSRaGuQ0Q/GTSJHimlCpRSp3X5fqFSqkopdbJSSiul3up2/DNKqd+3/3tW+zEPdjvmU6XUlQNRvxAiNNqzo0kpVa+UKlFKPamUSg51Xf3VnmkN7Z+rXilVPcDvLwPEEJJBkzBMKXUF8CBwFrCvvXmmUur4Hn6sAbhcKZUf3OqEEGHobK11MnAkcBRwe4jrCZQjtNbJ7V/p/v6wUsoSjKJE8MmgSRiilFoE/BM4Q2u9pstLfwf+3MOPVgNPAncErzohRDjTWpcA7+EaPAGglDpLKfWlUqpWKVXYMUPd/lp++4zOFUqp/UqpcqXUr7u8ntA+c1WllNoGTOv6fkqpSUqpVUqpaqXUVqXUOV1ee1Ip9ZBS6p32maLVSqkhSql72/v7Vil1VF8+p1LqJ0qpXUqpSqXUG0qpvC6vaaXU/ymldgI729smKqVWth+/XSm1oMvxZyqltiml6pRSB5RSP1dKJQHvAHldZrryPAoRQSODJmHEdcCfgFO11l90e+1BYHzXS3he3Amcr5SaEKwChRDhSyk1DJgL7OrS3ABcDqTjmr2+Tik1v9uPngBMAE4FfqeUmtTefgcwpv3rDOCKLu9lBd4EVgCDgBuAZ7vlzwLgN0A20AJ8Bmxs//4l4J4+fMZTgL+2952Lazb++W6HzQdmAJPbB0ArgaXtdV4EPKSUmtJ+7GPANVrrFGAq8KHWugHX77G4y0xXsb+1ir6TQZMw4nTgc2CLl9eacQ2KfM42tZ9lPgz8MSjVCSHC1WtKqTqgECily4yz1nqV1nqL1tqptf4KeA44udvP/0Fr3aS13gxsBo5ob18A3Km1rtRaFwL/7vIzM4Fk4C6tdavW+kNgOa5BSYdXtdYbtNbNwKtAs9b6Ka21A3gB16XEnmxsn8WqVkp1vPclwONa641a6xZclyK/121pwl/ba24C5gEFWusntNZ2rfVG4GXggvZj23ANrlK11lXtr4sQk0GTMOJaYDzwX6WU8vL6o8BgpdTZPfTxN+AMpdQRPRwjhIgu89tnSmYBE3HN5ACglJqhlPpIKVWmlKrBlTPZ3X6+pMu/G3ENhgDycA3EOuzr8u88oFBr7ez2+tAu3x/q8u8mL9/3tmD9aK11evvXjV3et7MOrXU9UNHtfbvWPBKY0WXwVY1r4DWk/fXzgTOBfUqpj5VS3+ulJjEAZNAkjCjFNT1+IvBQ9xe11m3AH3BdwvM2qEJrXQHc236MECKGaK0/xrW28e4uzUuBN4DhWus0XLPRXvPDi4PA8C7fj+jy72JguFLK1O31A36W7a9iXAMhANovv2V1e1/d5d+FwMddBl/p7ZfbrgPQWq/XWp+L69Lda8AyL32IASaDJmFI+3XzU4A5Sql/eTnkacAGzOmhm3uA44BJPRwjhIhO9wKnK6U6FoOnAJVa62al1HTgYj/6WgbcrpTKaF8vdUOX19biWi91q1LKqpSaBZyN5/qiQFsKXKWUOlIpZQP+AqzVWhf4OH45rvWgl7XXaVVKTWtfxB6nlLpEKZXWflJaCzjaf+4QkKWUSgvy5xFeyKBJGNa+duAUXNfc/9rtNQeu9QqZPfx8La677XweI4SITlrrMuAp4LftTdcDf2xf8/Q7vptJMeIPuC6F7cW14PvpLu/TCpyDa8F0Oa7Z8cu11t/29zP0RGv9Aa7P9jKumbAxwMIejq8DZrcfU4zrUuTfcJ18AlwGFCilanFdury0/ee+xbX+a0/7ZT25e24AKa1lpk8IIYQQojcy0ySEEEIIYYAMmoQQQgghDJBBkxBCCCGEATJoEkIIIYQwQAZNQgghhBAGBOVJy89vXSK35AkRQxZOucLopoRhT/JLiNgyLnMix+TOMJRhMtMkhBBCCGGADJqEEEIIIQyQQZMQQgghhAEyaBJCCCGEMCAoC8GFEP5TWpFEKjaTDWX4Ye8DR6NpcbbQQC1ayVppIcR3wj2/IDAZJoMmIcJEEqmkJqaCSROWmaPB5rRBI9RTE+pqhBBhJOzzCwKSYXJ5TogwYTPZwjtwFGDSrjqFEKKLsM8vCEiGyaBJiDChUOEdOICrxHAvUggx0CIiv6DfGSaDJiFEp7Wr1nHpKVdy8cmX8+xDz4W6HCGE8EuwM0wGTUIIABwOB/f+7n7+/uRfWLLyMT544yMKdu4LdVlCCGHIQGSYLAQXIgJde8HPqK5u9GhPT0/k4Zf+2ac+v9m0naEj88gbkQfAKWfP4tMVq8kfN7JftQohRFfByC8YmAyTQZMQEai6upHx197r0b7j4Zv73Gf5oXIG5Q3q/D4nN4dvNn3b5/6EEMKbYOQXDEyGyeU5IQQAWnvZtyQSFnYKIQQDk2EyaBJCAJAzJIfS4tLO78sOlpE9KCuEFQkhhHEDkWEyaBJCADDxiAkUFRzgYOFB2lrb+PDNVRx/+nGhLksIIQwZiAyTNU1CCAAsFjM3//EGfn75bTgdTs5cMIdR4/NDXZYQQhgyEBkmgyYhIlB6eqLXRZPp6Yn96nfm92cw8/sz+tWHEEL0JFj5BcHPMBk0CRGB+nNbrhBChFIk51eva5qUUhOUUpu6fNUqpfp3X6AQQgwAyS8hRCD1OtOktd4OHAmglDIDB4BXg1yXEEL0m+SXECKQ/L177lRgt9Zanq0ghIg0kl9CiH7xd03TQsDrE/CUUouARQBX33Elp/5wVv8qE73a9OlXvLtsBWXF5eTkZTNnwWyOPOHwUJclRLiS/Aojkl8iEhkeNCml4oBzgNu9va61XgwsBnh+6xIv23KKQNr06Ve88MQy8ufnkp8/iZqCel54YhmABI8Q3Uh+hRfJLxGp/Lk8NxfYqLU+FKxihHHvLltB/vxcMsakYjKbyBiTSv78XN5dtiLUpYkIdtcv/sG5x1zAlbOvDnUpgSb5FUYkv0QwDER++TNouggfU9ti4JUVl5OWn+zWlpafTFlxeYgqEtFg7gVn8I8lfw11GcEg+RVGJL9EMAxEfhkaNCmlEoHTgVeCWo0wLCcvm5qCere2moJ6cvKyQ1SRCIXqyhru+MmvqKmqCUh/R8w4nJS0lID0FS4kv8KP5JeAyMwvQ2uatNaNgDy5M4zMWTDbtQZgvusMraagnoLXDnLhVQu8Hh8riy5j5XN2WPni29gLd7Bi2dv88JqLQl1OWJL8Cj/+5hfExt92LHzGriIxv2RH8AjV8Yf07rIV7CguJCcvmwuvWuD1DyxWFl3GyufsUF1Zw/q3V/LQ+blcv3wlsxecSVpGWqjLEqJX/uQXxMbfdix8xq4iNb9k0BTBjjzhcEN/TF0XXQKu/53vau8ppCLtjKcvnzOSrXzxbc4eqxg3OJ6zxzZG1NmaEEbzC/z/25b8Cn+Rml/+bm4pIpC/iy47zngyTk9g2h2TyDg9gReeWMamT78aiHL7LJYWl3acpV1yjCtgLzkmlfVvrwzY2gAhwok/f9uSX+EvkvNLBk0xwN9Fl5F6O3AsLS7tOEvLSnZNFmclWzh7rGLFsrf71e8fbriT68+7kf17Crlg5kLeeuGdQJQrRL/487ct+RX+Ijm/5PJcDPB30WVZcTn5+ZPc2tLyk9lRXDgQ5fZZXxaXRqrNazbyUXEzz31V7NaeWb6xX1Pcd9z/6/6WJkTA+fO3LfkV/iI5v2TQFAP8XXTZccbTcW0dIuOMx9/PGcn+vOQfoS5BiAHjz9+25Ff4i+T8kkFTjPBn0WUsnfEIISKD0QyT/BLBJIMm4SFSz3hi7ZZdIYQnyS8RTDJoEl75MzMVLmLtll0hhHeSXyJYZNAU4yJxPxNfInUBqBCibyS/xECTQVMMi7bp4EhdACqE8J/klwgF2acphkXqfia+zFkwm4LXDlK1uxanw0nV7loKXjvInAWzQ11axCgtLuWmhT/jslN/xBWn/5iXHpdn3IrwJPkluhuI/JKZphgWbdPBkboANJyYLWb+7zfXMn7qOBrrG/nJ2ddx7InHkD9uZKhLE8KN5JfobiDySwZNMSwap4MjcQFoX32+ah0vL32Zg4Ul5A4fwvkXn8/MWdP71WfWoCyyBmUBkJicyMgxIygrKZdBkwg7kl+RLVLzSwZNMcLbgsme9jMJlwWW4VJHuPl81ToefWQx+efmMWLUVKr31vHoI4sB+h08HQ4WlrBz2y4mHzkxIP0J0R/ds2D8lHGsf219WOeXt7olwyI7v2TQFAN8LZi88KoFXHjVAo/pYCAsFlhG20LPQHp56cvkn5tH5tg0ANf/nutqD0ToNDY08bvr/sANv7uepJSkfvcnRH94y4L1r61n2vRp7Fi5Myzzy1fdkmGRnV+GBk1KqXTgv8BUQAM/0lp/FtBKRND0tP/Hbf/+uccf71033u3XfiG+zqT6e4Yl+5b4drCwhBGjprq1pY9KYXvhvn73bW+z87trf89p80/lpDkn9ru/UJP8iny+smDHyp3c9u+fux3rb36BZNhAi+T8MjrTdB/wrtb6AqVUHJAY8EpE0Pi7YNKf432dSe3Zupf169b36wwr2hZ6BlLu8CFU763rPFMDqN5bR+7wIf3qV2vN3355NyPHjuTCqy/ob5nhQvIrwvmTBf7mhmTYwIvk/Op1ywGlVCpwEvBYe1GtWuvqoFQjgqJjwWRXPS2Y9Od4X7f9rnztg37fDuxv3bHk/IvPp+D1Yip31eB0OKncVUPB68Wcf/H5/ep3yxdfs+KV99n42Zf8eO41/HjuNXz+0doAVT3wJL+igz9Z4G9uSIYNvEjOLyMzTaOBMuAJpdQRwAbgJq11Q9eDlFKLgEUAV99xJaf+cFZACxV95+8DLOcsmM3T/3mGrFkp2HLMtJQ5qFhVx2XXXepxbFlxOQnVg/j8n5tpLGsmMSeekbPyaG5oJi0/2e1Yf8+w5MGbvnVc93956ctsL9xH7vAh/OSaRf1eD3D4tMP4uOD9QJQYLiS/ooA/WeBPfoFkWChEcn4ZGTRZgKOBG7TWa5VS9wG3Ab/tepDWejGwGOD5rUt0oAsVfdeX/T8czU5KPqqgpb4NW7IVU4v3/1RsNhs73iog/4IhJI9MoH5fEzteKsBstvT7dmDZt6RnM2dND9idJlFM8isK+JsFRvMLJMNCJVLzy8igqQgo0lp3zHG9hCt0xAAJxC2r/uz/8e6yFSSMjKPs63ramhw4mh3kTM30unjRZFEMOSmDxDwbyqRIzLMx5PsZVK9soeC1g/0+w4qlfUtEUEh+hVigbrk3mgX+5BdIhgn/9Dpo0lqXKKUKlVITtNbbgVOBbcEvTUBoblnd/VUBpDgYfUUuqWMSqd3dyN7nSyivq/U4tqmhmdGHj6K+po6WtiYsVitDD8+jeuVer9sZSHiIgST5FVrhnl8gGSb8Y/TuuRuAZ9vvPNkDXBW8kkRXobhl1a5bGbdwKGkTXPtbpE1IYtTCIexcfMDj2Jy8bNrK7QwZM7izrWp3LTl52XKG5SeNdt0Qr0JdSQ90e52RRfIrRMI9v0AyLFAiIr+g3xlmaNCktd4EHNvndxF9FopbVrVTkzwiHpy47q90QvKIeLTT8z+0nhZdvvLI66x87QOaG5qJT4rn9Pmnct415wat7kjX4mzB5rSBSYdn8GjAqWhxtoRnfT5IfoVOuOcXSIYFStjnFwQkw2RH8DAXiucrJSQn0LC/lZTRCWi7E6VMNOxvJSE5wevx3hZdrn77M776ejOjLsslbWwyNbvqee/59wAkdHxooBYawWayocIwdTSaFmeLq04hDIiE/ALJsEAI9/yCwGSY0jrwU+1y90ngdF0T4G0xotEFlr7OmLwt0tyzdS/vvfUeoxZ+FxZ7nz/IGWedwegpo9yOr62oY+TCQW6hWLW7ls/v3syEa4aTMTHlu/Zv69j3dBnX/PpqeRZTlFk45YrwTMk+kPwKnEDlF3jPsO555G9+zVkwm3eXrSDj9ATDGbbn8RJGTx4l+RVFxmVO5JjcGYYyTGaawpyvW1bB+POVXnnkdVeIdDtjOrivhKJDhV6fSXcGZ7Dy6Q9obigkPimeM+a7Aqf7e+55ZA+Dq9Pc3i8tPxmH3Una2G57nIxNpqluvzyLSYgYEYj8Au8Z9u5z72F+1cKR/zehz/n1whPLqD3QQP5PjnR7P18ZlpQXT0tLCxmnJ0h+xSgZNEUAb4sR/Xm+0srXPmDUZbmdZ0wZE1NgIXzxyAZm/vwIn8+k6z4F7e0988/NZfc7+8k9JqfzuJqCeswWEzW76t3O0mp21aPMSp7FJEQM6W9+gfcMG7nAyZ4lxf3KL+bDpoe2e72E6C3Dqr6tI3FwvORXDJNBU4TyZ4Flc0Oz11kfh93pc8dbb1PhZcXlDMoeRcn+Q9jb2rBYrWSOSWPnM0VU7a51m34/9vhj+Or5zbAQtynyOEtcv3fZFUJENn8XiHvLsJT8eOwtDq99GM2v5OwULGaL1/2YvGXYvtdLmXBevuG6RfSRQVOE8meBZXxSvNdZH7PF5LUPR4vT6+W8OKeNA18Vk3lYKnG2BBwtDsq2VJCZk0HVyiaPvUxeeeR1jynyHVt3DvjCUCFEePF3gbi3DKsraMZiM7sd529+HfiqmLSsVC646jyv+zF1z7A0WxppQ1I83lPyK3bIoClC+fNMo9Pnn+q666PbrM+xxx9DwWuFHn00Njcw+irPy3nbHtiHfWUbGogfZKW5tI2DKyvIsGVx279/7vG+511zrscU+aZPv+LJfz2F02Z3u1Plyv93eVB+T0KI8OPvM9m8Zdi+ZYeIs9g8Zrn7kl++9mPqnmGSX0IGTRHKn2cadfzRd5/16Xr3XNc+7vvVA14v5zkdTpxtmpJVlbTV27EmW3C2aWrq/bt90xxvYtCsLLc9UYQQscPfZ7J5y7A587+7G07ySwwUGTRFMH92q/U26+OrD5+X8+JMjLpwiNdtBIx6d9kKxi8c+d20/HjIGlIrCymFiDH+7rbdU4Z1JfklgkkGTVGovw/I9HU5L84aR1yKBXuzA7PNjKPFQVyKBe3Q3HXj3YbeLxQ7BAshIsdA55fFbDH8npJfQgZNUSYQD8j0dTlvx9admBtNOKzOzgdbOqsUDhyG9y0JxQ7BQojIMND5ZW60kZaVavg9Jb+EKdQFiMDq+oBMk9nk2ktpfi7vLlvR777nLJhNyYpKbG0J5OXnYWtLYPfLBxh1xlDD7zdnwWwKXjtI1e5anA4nVbtrKXjtIHMWzO53fUKIyDbQ+VWyohKnXRt+T8kvITNNUSYQ08e+dhA/gzO48KoFbgsvTS0WRs8eZvj9/F0AKoSIHQOdXxdetYCn711qeO84yS8hg6YoE4jpY187iK98+gP+c82/3QLirhvv9vv9/F0AKoSIDQOdX+AaAPnznpJfsU0GTVFmzoLZPPq3x3DYHNgb2rAkWTG3mJl82GSum3ujxwN7vWluaMbR7GTr/XtpLm8lPjuOwSdm0tzQ7LFgcvyUcax+fjVZs1LcbsG97LpLB/iTCyEina/8+skvf+zzoePd+ZNfcxbMZs6C2Tz9n2ckw4QhMmiKMnu27qVNtzH0tCwSBsXRVNpK4etlbFi7kfE/HuY2XQ14DR2z2ULRe2WMvGAwySPiqd/fzL6XDqFQHgsmVz+/msaKZto+anXb7E0IIfzlLb8OLK9g+ZNvU1i63+OSG3hmmD/59cITy5g2fRqOZiclH1VIholeGfovQylVANQBDsCutT42mEVFk/7ePgsYPsMC19T06MtySR4VT5wZEnPj0Q5NycdVXqervfWTlpVKyslWEnNtoCAx18bgkzMoXFbu8cDLrFkptH3Uygm/Oabz56t2y74lInxIfvVdOOSXLTWOb/6zi0nXjTSUYf7kF/Nh5UMfcOT1E9wuz0mGCV/8GU5/X2tdHrRKolAgbp/1tagRvM8SNTc0k5yfgFlpTArMChIGx+Fo7PZgy7HJNDf4Xlw5ZGIOdeW1aKdGmRRDJuawz3nIY8GkLcdMS32be9+yb4kIP5JffgqX/ErOT0A7tNddvn1lmNH8SstPdj0MWB4iLgySOcgg6nr7LHx3ZuPPGUxPixq9hY4t0UbtrnpyJiYBYDFB06FWzIndHmy5q574pHjsbXaW3/sKQ+MsKOV6zdrsoOjVYlpNbaTFK2qaNTWOenSLZu+KA4yZO7yzn5YyB7Zkq3vfsm+JEBEvXPLr28cLcbQ62fH4fuKz4zqPbS5vxdQGX/z7Fbc+vOaXsx5TG2x9dDuJmbbOYxsrW3y2W5sdHn3vLKvhoj9dZeizi+hkdNCkgRVKKQ08orVe3P0ApdQiYBHA1Xdcyak/nBWwIiNVIG6fbW5oJm1sMlqDvdWOJc7S4xlW/shh7H5+L/ZzckgYHEfToVaK3izD3urk0LpK4gfH0XyolaLl5cyZfwafvvopF0wcygUnHdbZx6kT8rjq4aVkn55K4hAbuqSF8pW13L7wNP67ai1V31RBMlgcFiq21eJoc7LzjX3kTs+mrcHR44M3jdJa8826b7Hb7aA1B7fsw+zUALTU1DMoOZ6GplYarVbibFa3n4sfnE5aXqZbfyaTmckzJmIyydZkMUjyqw/CJb+qvq5ndFYmpYWNZByZ0tle+WUtty08jd9ecZpbH77y67aFp7Fs02aGz0qnLq6FlFYb+94s7WzPPyWDrPxkKgrq2fP6If59w/nMnTnZre8bnv7I8GcX0cnooOl4rXWxUmoQsFIp9a3W+pOuB7QH0WKA57cu0QGuMyIZvX3WYXfw0ZKVOJtbPTuxw/bH92PLtKKdTpTJREtlG9hhxYNv0OZ0csa187BYXf+nbKqux9Hs5MCKCuxNDiwJZhwtTrDDwfcraWuwY02yYFVWRk8ZxeHHTeXNh5fz6ZIPUO1TTdt3FVFZ3ED9i004Wp2Y40zYmzUTRw5m8fULuGXxKxzYXsXQQRk8f+tlOBxO/rT0PTa+9zVJiQnkjxlB6eYCVmwu8Ov31dLUQhpgs1lwOpycNGYIeZnJoGDcmceQnZ7cax8AX+8ppq6xxa2trLqe9//9Kmaz+6CptLYRW3oKon8WPnRFqEvoieRXH/hz+/+Otd9SsH4HqG4v9JJfDoeTyaceycgp+YD3/NIOqKhrwm5ysue5gzjbnJisJswOExuKqvh/T33o9pbe8qu1wUledhp/X3i2K79KXfl1z6LzmDtzMsd+PoJ7X1/FpyVFjBqSyd8Xnu0xYAJI0JoVD77R79+tCC8NMys55vIZho41NGjSWhe3/2+pUupVYDrwSc8/JeYsmO1aAzDfdYZWU1DvdRbmvcVv8euTpjBheI5HH6Pszfzz/VUMOi6NzLGJVO5qZP+LpfzqwtP47aWz+Nsra2hqaCalfUCRMDiNGZcOcQu6XasL2P9GCfFpNuyNDuLT4smZnNE5zT73+nPc3vOlH97G2DOyqNvfTENFG0lZVprr7fz+2bf53z03k2tRvLZoGNctb2T6pJFkpSUxc0o+19z1DItvv4ystKQg/DaNmzo6z2v7OcdNGeBKRDiQ/Oobo/kFsHv9dpZc9n2P9t7yq7qukV9+uKVz0OQtv6p217Lpoe2MOW0oZdsqqT/YSHJuEjmTMyjYV8fCG65xe09v+aXNij8vfZf1D/3CI78A5s6czLSJI1wZdpvvDPv75af09dcpwtkQ4wv+ex00KaWSAJPWuq7937OBP/a9uthhZPfYje9+wcmD0rwOmAB+e+Uc1mzexZrH97LLrrFaFMeNHsVvr5wD0LkOqYO3KfW6wnpamtsYcngGg09Lpbm0jcJPSnBWwb5v9wMwdExe52xVWXEFDWYbw8/JYfjweBoKm9mxuIiGlkqeemsN88aamDDIxryxzSxZvppbLpnNU2+toaqksPN7IcKB5Fff+bP7teoeRO16yy+A8kOVnTlUvOcgGWo0NQV1na9bkyw0VDdycFMZw+ZlkZg3iMbiVva/dgj7Ic2+b/eTlp1GenYa4D2/dj9zEEeF3Wd+AZJhwhAjM02DgVfb/ygswFKt9btBrSqK9LR77MG9JTRuK+DHP/b9B1peXU9LQw27bxhJdrKF8no7C5bVUFHT4PVsqOuUetXuWg58fJCSjVVkHZOKsoCz2o6yQMaURCo/rmHUzkKKy2rYuOcg0890TU8mpCYw4pyczsWbtok2hp/Zyv6XSnnizU/54EdZAFx+dBILlq1n3olHsvzj9fznvGyuW76eK+YdH/LZJiHaSX71Q393v+4tv9KSE/jREaNo2Ola45SbFE/DVxWkDEns7GPbqoOY40wMnZtF0gjXVgRqRDyWZBOO4mZG7Szk1VdXc+7tFwHe86t2WhM1n9Sw/OP1LFvgau/IryvmHY/WWjJMGNLrqlit9R6t9RHtX1O01ncORGHRzt5mZ/UT73LvFaf2eFzHmVF2smt8m51sYd5YE0uWr/Z6/Gk/OIWvH93Fhnu/punbaiZ8Lwdlh8HfSyfrsBRyjk4h67AUso5OJSU5ngUnHcY5Myfy/D0vsPXzbwCwmC3EpViwNztcCzibHaRPSsaabCUlro1/fFzF+U8UopRi3lgTv3zgReaOBnNrLXNHK7fayqvrOf+2h6moaQjQb04I4yS/Qqu3/FJKcdaMiSw46TAWnHQYd/3obJp2tpCcGcfIozNJzoyjuaQNm82GJckEbU7QGtqcmBNMpCTHc8rho1j76SbqqusB7/llSTBht2vmjXX9v7yu+bVk+WqeemuNZJgwJChbDqx65v1gdBtVynYXc/dFJ2OxmHs8btXGHRSXtrBECywYAAAgAElEQVR0S6lbe96hHZ1TyNXlNRRu20fh1gIch6q4/exTeHfj1xR8VkXaXjPjh+bQdrCVBCvUO8BqhpYDrYwb6rok+NqqjYxPbmXTU2+zd9VmEhMSULVWHFZNS1sTymmm9KNaHA12DikbL33dSIa1jWn3F5GaaKOpsZI/X5KOo62VM8dYuGHld2dqMuUtIo3kl//279jvtf3Z9zdyqKKGez6tcWvPz93sNQ86Fl93XZR9wmGTWL+rgANvl5GQYcHhBLMJGgubOXroUJ56aw3N1ZU88dv/cviMqcRbbZR9VAcJThwOJ2azicaCFmwWE0u3tPDAmmrSLa1Mu7+IzJQEsoq+oa25nvtPRzJM9Coog6a/njw1GN1GlcQzjsIW992t8q+s3sb+ijqP4045eZrPPp5Y+SXzjxnL8jVbOSI7lVvmHkN6imta+5YFJ3cet+ivz/DO+18z9AfZDBoRR+n+Vva9X87MCVMpr67n9Q/X8evjLdy57hBP/PIinshI4C+PrSTzqDRampuxOi1YqzQv//FHTJs4ggW33sd/5iVy3fJGvj/jCOLLNmNxNDMyw8K+6mbmjk5gyfLVXH7WcTLlLSKO5Jf/4k47wmv75BOO4bofz/Fo3/SE7yukc2dOdrtzrc3u4Oo7n+b9Xd8w+phksodZKS9qo+WbBoYlJ/P6h+tYfHYif1lXxC+OOodTkqz8/pV3GT4nhzprKyltcRTuL+eft13okV8v/uNmlixfjaNwvWSYMCQog6aM1MTeDxJuXtpayIlXnu7XzxQfrOBXr6zm+lmHccqRY3weV1JejbPcwerFxTi0xqwUKcpMSVY1T721hpOHtjIyxcHJea089dYafnHJbKbmD+bGh16ipqKGYYMzO2/NvefZFW4LKZ/6cAONTU08uqaV1HhFbbMGSxsTy3YA+Fx0KUS4kvwKHJvNSrKXQYbZ3PMMe1dWi5mq2nqocrLuqRK3DPvKWcg5o+2MS3dy2nA7r3+8kVsumU1Kos3r1gLd82vJ8tWs2riDb/fWSYYJQ2SnvzCRkGgjIyfdr68xh49h/q8u5pWyOp74YLPPvh//7ZWMHZzGigsSqLgxlRUXJDB2SBp337SA1z9cx5mjHIzMsHDmKAevf7SOipoGpk0cwfB4M19cM5xci2L6pJGUV9ez/OP1XH60KwQvPzqJtHgzg9OTWHNjPl/cMoY1N+YzLDuFu29a4HHs8o/Xy7oAIYTfvGXYqJwUEi34zK9ci2L9omE95tfyj9dz900LGJadIhkmDDH//ve/D3yvheuC0GnkKq+u57LfP85p0yaRGB/n9Zg3vipgzPQJXl/b9OlXPHn307z62Bt8+ekmklOSGTJiMAAms4kRk0ey8oNNTM1M9jp1/PDLqxjSupdT8xUVjQ6yk0yUNzh5ds0+pmfWcvQQuOGtRuaOt1Lf4mR7pWLzzkLy1QFGx9fiVBa+LnWyeWch460lnDrOdSaeGGdi76Ea6pvaOGdKMpc9e4BzpqbS1NLGY6t2c/qwVrdjK+pa2Vxi53uHjzH0OxERZPiMP4S6hICR/HLTn7/VN74qoLa1xSO/7CVVzD1ilOF+vGVYUVUrtY0tnD3e0uf8qqhr7cyqI3LjDGfYuBGDJb+iSfJgyMg3lGFK6yBsfrvmftlRt4t7nl3B8pUfM+/0k31O7V6y+F3O/oXnpnFdH5rZfYO5rrcCHywo4fW7l7Hqzis9+ph70318vbOAeDM0253EW0w0OwCTFZxtWJQmIx6qmsGuFePzh6PtTdx1YivJplbqnXHc9r844uKTKa+qcuu7tKqONgck2CykW1qptseRmZJAeW0z2anxHrXkDcrmjX/+1NDvRESQ427wvlFPJJL8ctOfv9Uz/vICO8r3eeTX4ZnDef03Fxnux1uG1bU4aXNCqk31Ob+AzqyqrGsynGGzjh4v+RVNhhwOo082lGHywN4g65gS9raQsLy6vnMX7QVHjGLpX5Yy9+bziE/87g+1t4dm7ty8m0d++R8unfs9Hlk012sNp0+fxPey61m1o5r/nJXAdW818f0J6WysTeeolGov7TZmpDeSoFoZmmJif20rs/OtxI2cxC2XzHarOystifLqeo/FlT0tluzpdyKECB9G88vX3+/WrTsZd81Ij/za9Mguw32A9wxLtFmZmG1i44GWfuVXRx1GM6zzWMmvmCRrmoLMfQda9/2Vut7Keu6MCdx/8cm888+XaKxv6jymrLictHz3Z66ljkhi/7eFvPrX5/j84dcYHdfMIKtmwsjBXmtYtXEHT26o5YjB4NBOjhgMT3xRy5fbC322/2dNFfOfb+SkJxuY/3wj/1lTxaqNOzzq7u0z+vs7EUKED6P55UtdXYNHfqXlJ1Nb12i4D/CeYWv3NfL0l/X9zq/ePqc/vxMR/WTQFES+Fh5W1DS4ncF1tA3KSOGhK05lxb9eprSoDPhuh2+AlppWij45yMb7tjIqO40HLz6ZpspSHv1hTo8LFB//7ZXkD0rl13OGM3lUHr+eM5z8QamsuP8Wr+3P33ktVpPmmR8ksvbqZJ75QSJWk+afNy/wqHvH/lK/Fkv29DsRQoQPf/PLm5y0ZLY+up1dy/Z0fm19dDvDB2cY7gO8Z9iE4VlMHJ7dr/zq+lmMZJLkl5CF4EH08MurvC483Fxi71yUePbkJLcF0onxcZw2eThvvvQptUD+pFF8/NQaDm0upfVgI8kJJih28NdL5/HB2q1e+zBaR8dCx8mDrPzwiQLOPSyNllY7f3nlS87Kb+PMcRbS401YzVDVpHlhXTHVtY1u79nbgm9/fifejhcRQhaCR52+5Fd34/Ky+WjjDkbPymbszBxs8Wbqvmrg75efbTi/fNXy5ldVTM/VHD0soc/51fWzGMkkya8oJQvBw8M5P3uA4tJyj/asjAzamutZtiCly/OY6jyuo//lpdWs33cIa0sLZZVlHKioZdSQTG4+d1bnJm299dFTHR0LHQ9W1JET76Cs2UxuVgp7S6qJM2myExUmBU4N5Y0ai9XGuLw0t/c8/oFCEhITsJjdJy07Fnwb/Z34Ol5ECFkIHnX6m18d3vl8G/e+voq9JZV9yi9ftRwor8NqBoeTPufXgmV1PheIe8skya8o5cdCcBk0hcA9z66AAxu45aS079o+qYGhxxi+EyMQfQBs33eIs274G68sSOS8ZY288+BtvPnJJq99/688lROza/v9niIKyaApZkh+iagjg6bwFoizld76MHpXyg9ufZAp5n1ceaSVJze1sdUxEodDU1xaTmubg8LyWoZnpxJnNfe6jYCIYTJoihmSXyLqyKBJGNlbpeMs7bUF8VjbHyA+f1kz7zx4G+OGD+IHtz7I/j27GDF6LK/+/f8G+BOIiCKDJhFAkl9iQPkxaJK756JQx0N4bzvO3PlYgY728297uPP72x58iYVTLFjNMDLdtWBy4RQLt97/Itv3HWLLt7u585R4tny7m52FpaH8SEKIGOErvzpe68gwyS8RCjJoikLdH8LbdT+lrvuTfLm9kMc2tnLmsw3M+G89Zz7bwGMbW/lyeyG3PfgSP5hgZlS64gcTzNx6/4uh/EhCiBjhK786XuvIMMkvEQqGB01KKbNS6kul1PJgFiT6p+MsrftDLDv2U+q6P8kXS37DpJE5rL15DJtvHc/am8cwaWQOz995DVu+3c15Ey2MTDdx3kSLnK2JiCb5FRl85Ze3vaFW3H+L5JcYcP7MNN0EfBOsQqJZ98tivbX3p4+Os7T8DDPxFhP5GWZOzmvllw+8yNzRYG6tZe5oxZLlqzt3ts1Odj1NJzvZwryxJq7729Ous7QME/EWxagMk5ytiUgn+dVH4ZBfHXnVNcN++cCLkl9iwBl69pxSahhwFnAncEtQK4pCXaeUuy5q9NXenz5WrvuGr3fWsXybCZMJnE4oqXeQbGvkz5dm4Ghr5cwxFm5Yub59f5IWlm5xPwPbX1LNM6Wad3e1ue1zYo3bH9hfjBADQPKrf0KdX2WNTsaXf422N3H/6XRm2H8/L2BvUQJLt7S4vZfklwgmow/svRe4FUgJYi1RydcDL/15aG3HlPWvjzNz50freuzj9OmTOH1oIxcdkchVS/ez5JIRXPtyKVNyFBZHMyMzLOyrbmbu6ATMwyd5Dbue9lARIgJJfvVROOTXs5sa+V+5jeMym9wy7OqZGZiHH+uRYZJfIph6vTynlJoHlGqtN/Ry3CKl1BdKqS8Wvy4PMOzg6+GO/j4g0tfC7u59rNq4g6VbWjj2vv2U1Dk4+t79fL6/mSe+aODspQ2c8Hg9Zy9t4NF1tZ0PsOyus48HSzu/lm5p8Xm8EOFK8qt/wiG/lm5p4cvthTy6rs5Qhkl+iWAyMtN0PHCOUupMIB5IVUo9o7W+tOtBWuvFwGJA9jlp13EmtWyB6wT38qOTWLBsPfNOPNJru7eztY6ztL+d0LEw0s4vP1rns48X/3Ez5dX1nHXD33jqvGTOW9bIs3+9kdvve9bj8QFP/O4qr3XLRm8iikh+9VG45Ndzf7mejJREr49d8ZZhkl8imHqdadJa3661Hqa1zgcWAh92Dxzhna+F1r4WMC5Zvtqvhd2++rjtwZe4eKqFw4dYuXiqhWvvesrnsUJEM8mvvutLfoH74u5A5Net97/osxbJMDHQjK5pEn2wauMOiks9F1qX1xZSeDDeoz3vkGv62MjCSJOlhcK0BI8+0vd9TUFhEQ/9KBmA66bHs2RTJU9uTPFYMJl3aIc8c0kI4VVf8uuWS2a7Le4ORH6d+PhuahtaqKn1rEUyTAw0eYxKGCmvrmf+Lf8iSTfQqJJ47V+3uM6k/Hiw5Q9ufZDDLPv44ynfTZP/7sMGtthHyqMERPDIY1QEnhk2+/ijSKrYIvklwpsfj1GRmaYw8tRba8ixNlPT0EZ2UnPnwkhvZ3u+zrC+3F7IutY2Hvuy2q3dGlcY1NqFEKJ7hr384QYsJi35JaKGDJrCRHl1Pa98sBZzSwuLz05i0ZuNvPrBWl771y09PuW7u4LX7wpilUII4Z23DEuxxfuVYZJfItzJs+fCRMcZ2rkTrEzINnPuBCvZ1uYBWejoz86+QgjhTagyTPJLDCQZNIWJ99Z+w+biJmYMU2wra2PGMMXm4iZWrA3+kx+6P8hXCCH8FaoMk/wSA0kGTX0QjDObM2ZM4qcn5nD85GFMHpXH8ZOH8dMTc5g9Y1JQa+n+EEw5WxMi+oVLhkl+iUgjg6Y+CMaZTV93se1vLf7s7CuEiA7hkmGSXyLSyEJwP/nzzCV/9GUX2/7W4mvH30B9JhH9DpRV09DUyvjjQl2JMCpcMkzyS0QimWnyUzid2fS3FtllV/Skuq6R/SWVnV8rNuzmD0tXc8cz/+OK+1dx7eNf8Jv3Krj/m9RQlyr8EC4ZJvklIpHMNPkhnM5sAlGLv3tAichTU99ES5u983unU/PaZzupa7a7HbentB5Tgvvgp85hJXlwfuf38anjGH3OTwAYaotHqejZzzJWhEuGSX6JSCWDJj/0dGbTlz/S8up6rrnrGRbffllnUHhrC1Yt8mDL8Od0Oqlvcn/8zY6iSj7fXuLWVt/Uwu4qBzabrbNNa6h2xJGcNcTt2CGTzydj0FC3tsNsCVi7/KyIToHMMF9ZZSTDJL9EpJJBkx8CfWbTdRFkx897axuIWkRwOZ1O7A5n5/frdxzk28Jyt2O2FdXSZo53a6uqayIuZ6TbrE5ccib5085xazMrxbTMHJn9ET0KZG74yiojGRYL+VVaVceba3cD8lSecDfh6AROGH2yoWNl0OSHQJ7ZeFsEqbU2vDBSzrLC3+ffFLG7uJLPdlVQ1ayIy8jtfC158AhyJ5/pdvyImRkkJqcMdJkihgQqN3wt4ja6uDsW8uvlNTvZM+QMMnKG9H6wCKncvNzeD2ong6YQcV8E+d2uud3bouWsKxa8v2E3X+wupajGTllrHNmjpjJo/ClMnDEMW3xCqMsTImC85dctl8z22R6TNGTnDiNzUF6oKxG9SM80fjOLDJpCwNsiyPOfW4dTa169OK2zTW6fDV9FpVVs2FHM+9vKqG92YE/IJHPcMQyeOZ8j8kaEujwhgsbXIu55Jx4ZFovMhQgmGTSFgLdFkCcPbWXLIQfZyVmdbf1ZZC4Cx+l08tnXBewvr2ftnipKGiAhayiDp57K5MsOw2KNC3WJQgwYX4u4f/nAiwG9USYcPfvRNjYWVBo6dldhKd8/LjHIFYmBJoOmEPC2CLK0qoE2Bxz7YPQujIwUjc2tbNldzPINheyvbEIlZZI8ZhrJmYOZfOLRTJHF1iKG+VrEXV5bSOHB+Khe3L12bw2HXf5XQ8dODHItIjR6HTQppeKBTwBb+/Evaa3vCHZh0SwWFkFGkt1FZazdUcK6XWVUtMWhrAlkTpzJ8DkXMiIzO9TliX6Q/Aq8WM4vuTtVGJlpagFO0VrXK6WswKdKqXe01p8HuTbRhdH9m0TP6htbKCip4IXVuyiracIRn4kzNY9BY2Yx5UczQ12eCDzJrzAg+SWiRa+DJq21Burbv7W2f8nGEwPM6P5Nwt2BsmqKymt598tC9la04IhLIS3/MMaedTvDklMxmeRJQtFM8is8SH6JaGFoTZNSygxsAMYCD2qt1wa1KuEmWA/YjDZOp5Oy6nreWLubLfsraVKJ2FPySMwYxMR5lzEkPlEGSTFI8iu0JL9ENDE0aNJaO4AjlVLpwKtKqala66+7HqOUWgQsAnjk1gtZdO7xAS82VsneJ941tbTy1Z5SVn9zgG0Hm2nVCtvgseRNOZfxJ4zBliB3rgjJr1CLxPz67JsiNuw85NFedKiSqSGoR4QPv+6e01pXK6VWAXOAr7u9thhYDMCa+2X6O0DC5QGboeZ0OmlsbuOV1d9SVFFPUb2ius1KxshJ5B83n2mZOTKLJHok+TXwIjW/nv7fHkadc4tH+8kny4lYrDNy91wO0NYeOAnAacDfgl6ZAAL/kOBIUFBSyZptB3A6NZ/tqsBiS6CitglzRh7Dj55H6pQcjhpkfNt7Ebskv0IrUvPLarWSnj041GWIMGRkpikXWNK+LsAELNNaLw9uWaJDtD7YsqCkkrXfHOB/O8qxxsVT19hMoyUdi9WCOT6F/GnngVJM/t5w4mzxvXcohHeSXyEUrfklYpeRu+e+Ao4agFqEF9G0J0ptQxPLP9/Bh9vKaE0Zxojp85gyc7zsqC2CRvIrtKIpv4QA2RFcBJnD4eSTTbtY8skeHMlDGDptDtOvPVo2iRNCCBFxZNAkguKTzXtY+uluanUSmVNP4tifXIstPiHUZQkhhBB9JoMmERBaazbtLOK5/+2iqF6ROvoojrr6ekxmc6hLE0IIIQJCBk2iX/aXVPLCp9vZsLeGtMknMPn8OxiXnBLqsoQQQoiAk0GT8FtVbSNvrt3J+9sqMOeMZtzxV3PKuSNlnZIQIuSu+Nc7JKYP6lcfbXHyoG7hnQyahCFNLa0s/Wgbn+2qoDUhh+HHnMoJ10+XgZIQIqyotKEcfdntoS5DRCkZNAmfWlrbeP/LAlZsLqLCHs/ok3/IMbMmY7FYQ12aECLG2e0OGlta3dq0BqfDEaKKRCyQQZNw43Q62by7hFc+283uOgt5h5/I2IU/YWpi+D7yQIhAam2zh7oEYcAvHv+YhrQxbm1aa8aeND9EFYlYIIMmAUBJRS3PffINawsayBhzJGPn/JzvZ8p1fRF7rnlud6hLEAYkDZvGjDMuCnUZIsbIoCmGNbe08eDyjWwuqkdlDGfsCVdy6tmjQ12WECF13MWeD2oVQgiQQVPMcd35toOPvy2nyZrOuO9fzInnTgx1WUIIIUTYk0FTlCsqrWJvcSXbD9awpaiWkpZ4Rp90PjNmHSF3vgkhhBB+CPqgyeFw8uWOIjTarX1zQSXbS+rd2rTWHKxzYMVOTqrnk+2dDgdzj8wlLcnztZTEeCaOHBzY4iNEc0sbW/YUU9/UyvKNBzCZzZTUtmJLycIZl0zGuGNJGJnG1NOnMjXUxQohhBARKiiDpkVPfNn575bWNlLGHIu123PHkobnMP70Izx+dkIP/drtbbz1xf+g2enxWv22Qvhgo9fZk9bacgal2jzanU7NGUfkMSTD884wm9XC+BH92yAt0HYWltLU0sbyL/ZR3WSnsrYJlZJDS5uD9AkzMVusjLv4OqxxNsaHulghhBAiygRl0DTtst8Eo1ssFiuTZ54SsP6cTidvf/4+ztpWj9caqspQK77AbDZ5vNZSW0luepxHu9Zw8pRcxgxJ83jNpBSjh2b3ekmsoqaBytoGauqbeXVdAS1tdspa4oizxaPTh5OUmUfeCWcybsgwPz6pEEIIIforptc0mUwmJh832++f01qjnZ6zXVprVqz/kLZttR6vtTTU4ShbS1yc56+8ub4GNMSnpNGobaQOHQumdCacfyVWaxwT5KG3QgghRMj1OmhSSg0HngKGAE5gsdb6vmAXFs6UUigfA5lJ3/N/ENaxg61JBkdCBJTklxAikIzMNNmBn2mtNyqlUoANSqmVWuttQa4tZshgSYigkfwSQgSM54KdbrTWB7XWG9v/XQd8AwwNdmFCCNFfkl9CiEDya02TUiofOApYG4xiRPT7608vor6+zqM9OTmF2x94Lmz7FpFP8kv0l+SXMDxoUkolAy8DN2utPVY6K6UWAYsALv3ZnznpHHkmkPBUX1/H6Kvv92jf898bwrpvEdkkv0QgSH4JQ4MmpZQVV+A8q7V+xdsxWuvFwGKARz/Zo70dI4QQA03ySwgRKL2uaVKujYUeA77RWt8T/JKEECIwJL+EEIHU66AJOB64DDhFKbWp/evMINclhBCBIPklhAiYXi/Paa0/BeTJrsIvvhY1VpYexFaw06O9pqK83+9ZU1HOAS99V5Ye5NdXzvNolwWW0U/yS/SVtwyrLC3BsmsbZovVrV3yK3bE9I7gInh8LWqs+NMPqFjueZVEO+39fk/ttHvv2+GQBZZCCL94y7DaB39K6Yt/IC4l061d8it2yKBJDCiTNY7Df/ofj/ZABEB6zhCv4bLhrgv73bcQQuScdRMVy+/xyDDJr9ghgybhwdeltbrKMlIyc9zaairK0U476TlD3Nqry0q89u1sa2HjvVd7tNtry/o9BV1dVsJXD1zn5T09H8gshIhO/uRXT+3eMqx0+b046io8MkzyK3bIoEl48HVpbcNdF3q0HyjYScXyezzafZ4dmSzk/egBj+bC+y/t9xS0VibyrrzXo33vvy8x3IcQIrL5k1+9tXen7S0MuvDPxGWPcGuX/IodMmgShjlavc8SORqqDPehHW0UPXi5t1e8LoL0tsDyujOORHt5Xp+222kq3e+la+Pb7gRqV17Z3VeI8OIzv1pbDPdhry2n9IXfeHnFeH6BjwzTUPjkLWSfeXO3dv+2DQtE9kh++SaDJmGYMnvOEjnbWihe/BOPY03K5PUsS5mtDL3uSY/2ogcvx9bt7A3AqZ0ebdpsZvhPn/FoL7z/Uizpg71V7rWW5OQUj7ZA7coru/sKEV685Re4csMbbxmmzJZ+5xd4z7DW8v2Uv/F3LxlmPL8gMNkj+eWbDJpEUKRlZXPnk8s92q+dewzKZGR7sL6xxtk82kxms9dahBDCF28ZFuz8As8Mk/wKLzJoimG+pmDLiwup9jKNrR2et9WWvngH2un0WMDYUuV9bxHtsONs8z4d3trS7Pa9vaYUp8PB9fOmd+tk4J9yUV1WInulCBFGApFfHe3eFmD7yrCSpbcx6Id/8Gjvnl8ATofdM79gwDNs639/RnN5qcfnkfzynwyaYpjPvZTuPJ/sc271aD/03K9oLXdfM+Rsrmfwgj9itrpv9nZw6a989H0Bymz1aNcOO4eW/tKtzVFfiSVtMMOvdN+7ZO99vhZGaoqfvNmjVfmYIveHViaZrhYijAQiv8B1yS1r3i0e7d4yzLJzGxXv/Nsjw7zll6vdwciblnq0e80wpXA0VHtkWCDyy9HcyJCFf2Zo/ji3dskv/8mgKQQid5Gd9rhrxNns+hwOu/tZnO7hTMrX9LajvtK9D6cT8OeMTHndA+rLv10ks0RCBEg05VdHe/f8gh4yTHnPsO755S9L6mCUyeSRYZJf4UUGTSEQ/ovslI9wURQ+0G3RpAZLZh4ma7zBvrXXu+eU2Uzeokfd2uw1hyh/4++ePTjsnnUAONq8/g610274952cnOK13aT8W8fgqx9fizeFiBRRlV8dxw8aZahnjcZeW+6RYd7yC+DAwz/23o+3DNOuu4u7/x79yS/wnj1t9ZWYvdxx7Ivkl28yaBJ+efjtDW7fX3vmMX4MmAClyL/haY/mvfddgsnquYjbG3OcjYeWrzP8lt7O0nzxdebmTx899SOECJ3u+QWuDDNKKROW1GyGX3WfW7s/+QX+ZVggsufXV85jyPDR/epDuMigSfSPjtwda2sqymXaW4hYF6EZJvkVGjJoimG+pmC1w07x4z/1/AEvD6XUTjuHlnouuvQ2zQygHA72PeBtc0u83FXnWhjZ/XiLKTAPrXdqZ78v28l0tRChEYj8Av8yrLL0INpHhnm/K1h7PTYQGSb5FRoyaIphvs5Gfnr2TAbN97wTpPSF33m0KbOF3Cs8t/4vfOBSv/YWuX7edOJs7pf54gaNwGQ2+3UpLljkzE2I8BKI/ILAZJi3/AIwmS2SX1FGBk0hEO6jfu20U7H8Hq/tHhwO72d1Dodf72kxqaCdkfn6fQfiVl4hYk1U5RcEJMMkv2JHr4MmpdTjwDygVGs9NfglRb9wH/Wn5wwxPO2blTssIHfS/PuNtX4d749ALe4WkUkyLLCiKb8gMBkm+RU7jMw0PQk8ADwV3FKEcBe5+8GIMPMkkmEiBCTDok+vgyat9SdKqfzglyIiUTCn6oO9H0y4X2YQgSEZJnoSqRkm+RUasqZJ9Eskny1FcnaMhY8AACAASURBVO1CiMCI1ByI1LojXcAGTUqpRcAigEt/9mdOOueiQHUtBpicwYhYI/kVPSS/RDAFbNCktV4MLAZ49JM9A/8YehEwcgYjYo3kV/SQ/BLB5N8DtYQQQgghYpSRLQeeA2YB2UqpIuAOrfVjwS5MCJlmF4EgGSZCRTIs+iitAz8TLdPbQsSWn5w0OjDPtgkDkl9CxJYpeakcNzbbUIbJ5TkhhBBCCANk0CSEEEIIYYAMmoQQQgghDJBBkxBCCCGEATJoEkIIIYQwQAZNQgghhBAGyKBJCCGEEMIAGTQJIYQQQhgggyYhhBBCCANk0CSEEEIIYYAMmoQQQgghDJBBkxBCCCGEATJoEkIIIYQwQAZNQgghhBAGyKBJCCGEEMIAGTQJIYQQQhhgaNCklJqjlNqulNqllLot2EUJIUSgSH4JIQKl10GTUsoMPAjMBSYDFymlJge7MCGE6C/JLyFEIBmZaZoO7NJa79FatwLPA+cGtywhhAgIyS8hRMBYDBwzFCjs8n0RMKOnH5iYm9KfmoQQIlAkv4QQPRqcFm/4WCODJuWlTXscpNQiYFH7t89orS8zXEUEUkot0lovDnUdwSafM3rEwmf0QvLLh1j47yEWPiPI5xxIRi7PFQHDu3w/DCjufpDWerHW+lit9bHApADVF84W9X5IVJDPGT1i4TN2J/nlWyz89xALnxHkcw4YI4Om9cA4pdQopVQcsBB4I7hlCSFEQEh+CSECptfLc1pru1Lqp8B7gBl4XGu9NeiVCSFEP0l+CSECyciaJrTWbwNv+9Fv1F9bJTY+I8jnjCax8Bk9SH75FAufMxY+I8jnHDBKa481kUIIIYQQoht5jIoQQgghhAEBHTQppR5XSpUqpb4OZL/hRCk1XCn1kVLqG6XUVqXUTaGuKdCUUvFKqXVKqc3tn/EPoa4pmJRSZqXUl0qp5aGuJViUUgVKqS1KqU1KqS9CXU84kvyKHrGUYZJfA1xLIC/PKaVOAuqBp7TWUwPWcRhRSuUCuVrrjUqpFGADMF9rvS3EpQWMUkoBSVrreqWUFfgUuElr/XmISwsKpdQtwLFAqtZ6XqjrCQalVAFwrNa6PNS1hCvJr+gRSxkm+TWwAjrTpLX+BKgMZJ/hRmt9UGu9sf3fdcA3uHYdjhrapb79W2v7V1QuflNKDQPOAv4b6lpEaEl+RY9YyTDJr4Ena5r6QSmVDxwFrA1tJYHXPuW7CSgFVmqto+4ztrsXuBVwhrqQINPACqXUhvbdr0WMi+b8gpjJMMmvASaDpj5SSiUDLwM3a61rQ11PoGmtHVrrI3HtoDxdKRV1lyuUUvOAUq31hlDXMgCO11ofDcwF/q/9UpSIUdGeXxD9GSb5FRoyaOqD9mvkLwPPaq1fCXU9waS1rgZWAXNCXEowHA+c0369/HngFKXUM6EtKTi01sXt/1sKvApMD21FIlRiKb8gqjNM8isEZNDkp/YFho8B32it7wl1PcGglMpRSqW3/zsBOA34NrRVBZ7W+nat9TCtdT6ux2t8qLW+NMRlBZxSKql90S9KqSRgNhC1d4gJ32IhvyA2MkzyKzQCveXAc8BnwASlVJFS6seB7D9MHA9chmtUv6n968xQFxVgucBHSqmvcD27a6XWOmpvZ40Bg4FPlVKbgXXAW1rrd0NcU9iR/IoqkmHRI6zyS3YEF0IIIYQwQC7PCSGEEEIYIIMmIYQQQggDZNAkhBBCCGGADJqEEEIIIQyQQZMQQgghhAEyaBJCCCGEMEAGTUIIIYQQBsigKYYppX6llJKnYwshwpZS6kml1J99vPawUuq3oa4jQP3XK6VGt/87QSn1plKqRin1olLqEqXUimC9tzBOBk0RTClVoJQ61L61fEfb1UqpVUZ+Xmv9F6311UGoa5VSqrk9BGqUUp8opQ4L9PsIIUJLKXWCUmpN+995pVJqtVJqmlLqSqXUp8F+f631tVrrPwWiL+Vyo1Lqa6VUQ/uu8C8OVHZprZO11nvav70A107YWVrrH2qtn9Vazx6IOkTPZNAU+SzATaEuwoufaq2TgSxcD8t8OrTlCCECSSmVCiwH7gcy/z979x3fVnX3cfxzNCx52/G243hk7zAyGlYYCQQSQoGGVSil7LIeygN0PC0dtNDSlllKaEvDCBBGEkhCSAiEkZDByIZsJ0684ykP2ZLO84dkY8dyInlJsn/v1yuv2FdXVz+J6Mu55557DpAB/BawB7KuLngCd5behfv9DAMWAxcFoJYsYLfW2tHVAymljN1Qj/CQRlPo+wtwX/PilMdSSj2hlMpXSlUrpb5USp3R6rGHmlfFVkqtUErdccxztyilLvX8PEIptcpzNrlLKTXXl+I8X/rXgFGtjjtJKfW5UqpSKVWolHpaKRXmeewZpdRfj6njXaXUPZ6f05VSbymlSpVSB5RSdx1z3C8877VYKdVnFyQVIggMA9Bav6q1dmqt67XWK4Em4J/A9zy9zZUASqmLlFJfe76f+Uqph1ofrFWvVaXn8euPfUGlVLRS6iOl1JOenqGWS2ZKqWme3qGfKaVKPNny41bPTfBkSbVSapNS6g/NvWFKqaHAT4GrtNYfaq3tWus6Tw/PI17qiFdKLfXkUIXn54GtHr9eKbVfKVXjyalrPNuHKKU+9vTMlSmlXm/1HO15/LfAr4ErPJ/fT47tuTteHns+k2eVUsuVUrXA2b7/JxUnIo2m0PcF7p6c+zp4fBMwAfeZ0wLgDaWU1ct+C4Crmn9RSo3CfbazTLkv/63y7JPs2e8fSqnRJyrO0xi6BljfarMT+B8gEfgecC5wu+ex+cBVSimD5/mJnsdf9Wx7F9iC+6z2XOAepdT5nuc+ATyhtY4BBgMLT1SfEKLTdgNOpdR8pdRMpVQ8gNb6G+BW4HPPJafmE7pa4DogDnfvzW1KqUsAlFKDgPdw91ol4c6sza1fTCmVAKwG1mqt79LeF05NBWJx58NPgGea6wKe8dSQCvzI86fZucBhrfVGH9+7AXgBd0YOAuqBpz11RgJPAjO11tHA1Fbv5ffASiAeGOh5v21orX8D/BF43fP5/fuYz8GXPL4aeBiIBnr8Mml/Io2mvuHXwJ1KqaRjH9Bav6y1Pqq1dmit/wpYgOFejrEImKCUyvL8fg3wttbaDswC8rTWL3iO8xXwFu7r7h150nOGaQPuwN1t31zTl1rr9Z5j5QHPAWd5HtsIVOEOMYArgTVa62JgIpCktf6d1rrRc/3/ec8+4D7DHaKUStRa27TWrRtqQohupLWuBk4HNO7vYalS6h2lVEoH+6/RWm/TWru01luBV/F873HnzQeeXqsmT2a1bjSlAx8Db2itf3WcspqA33mOsRx3/gxX7ktUlwG/8fQg7cR9gtYsASj0470f1Vq/5TlWDe4GylmtdnEBY5RS4VrrQq31jlb1ZQHpWusGrXVnGjS+5PESrfVaz2fd0InXEB2QRlMfoLXejntswYPHPubpqv7G0x1cifssLNHLMWqAZXzXALkSeMXzcxYw2dNtXuk5zjW4z9g6cpfnDNOK+0v+plJqnKemYZ7u7CKlVDXus6rWNc0Hfuj5+Yd8Nx4qC0g/po5f4B4wCe4zy2HAt57u91nHqU8I0UVa62+01tdrrQcCY3A3bh73tq9SarLn0lqpUqoKd29U8/c+E9h3nJe6CAjHfdnveI4eMw6oDojC3XtlAvJbPdb656NA2gmO3UIpFaGUek4pddCTYZ8AcUopo9a6FrgC9/srVEotU0qN8Dz1fkABG5VSO5RSN/j6mq34ksf53p8qukoaTX3Hb4CbcHdLA6Dc45ceAOYC8Z5GTBXuL603r+K+NPY93AH1kWd7PvCx1jqu1Z8orfVtJyrKc6bzKbAXaL7741ngW2Co51LaL46p6WVgjlJqPDAS92DM5joOHFNHtNb6Qs9r7dFaX4W7y/pR3A21SIQQPU5r/S3wX9yNJ2+XzhYA7wCZWutY3A2g5u99Pu5L6h15HlgBLO/kd7oUcOC+JNYss9XPq4GBSqlTfTzez3D32E/2ZNiZnu0KQGv9vtZ6Ou6G2Lee+tFaF2mtb9JapwO34L6sNsTP9+JLHnv7/EU3kEZTH6G13gu8jvvOj2bRuIOiFDAppX4NxBznMMtxn8X8Dvf1dJdn+1JgmFLqWqWU2fNnolJqpC+1eRpho4DmLupooBqwec7A2jS+tNaHcY/Fegl4S2td73loI1CtlHpAuecxMSqlxiilJnpe54dKqSRP3ZWe5zh9qVEI4R/PYOSfNQ+AVkpl4h5fsx4oxt0ICWv1lGigXGvdoJSahHvcTbNXgPOUUnOVUiblHrQ94ZiXvAPYBSxVSoX7U6vW2gm8DTzk6SUagXt8VfPje4B/4B47OU0pFaaUsiqlrlRKtevB97yXeqBSKTUA90lr8+eSopS62NO4s+O+ROj0PPYD9d2A8QrcjRt/M6pLeSy6RhpNfcvvgNZnYe/jHly5GzgINHCcblvP+KW3gfNwnxU2b6/B3Ut0JVAAFOHuybEcp5anPXd+2HA3fn6ltX7P89h9uAOzBvcZ2Otenj8fGEurqQo8wTcb9yDRA0AZ8C/clxwBLgB2eF7zCeBKuZ4vRI+pASYDGzx3aa0HtuPuhfkQ90lSkVKqzLP/7cDvlFI1uMdhttyoobU+BFzoeW457oHT41u/mGfg9824M2xJBze0HM8duLOiCHeuvErb6RHuwj2Y+xncJ137gO/jvvnkWI/j7o0v87zvFa0eM3jeR4HnvZzFdze6TMT9edlw97rdrbU+4M+b6GQei26ivN+AIERgKaXOxH2ZLrtVj5cQQnQLpdSjQKrW+kcn3FkID+lpEkFHKWXGPcncv6TBJIToDp7LieOU2yTcN44sCnRdIrRIo0kEFc91+UrcAyi93oUjhBCdEI17+EEt7kuDfwWWBLQiEXLk8pwQQgghhA+kp0kIIYQQwgfSaBJCCCGE8IGpJw76zu435ZqfEP3IxcMu72jC1JAj+SVE/5ITN4SxyRN8yrAeaTTVNdX2xGGFEKLHSX4J0b80Ou0n3slDLs8JIYQQQvhAGk1CCCGEED6QRpMQQgghhA96ZEyTEMJ/SisiicFisKAIvnHVGo3dZaeWarSSsdJCiO8Ee35B92SYNJqECBKRxBATEQMGTVBmjgaLywJ1YKMq0NUIIYJI0OcXdEuGyeU5IYKExWAJ7sBRgEG76xRCiFaCPr+gWzJMGk1CBAmFCu7AAdwlBnuRQojeFhL5BV3OMGk0CSFabFizkR+ecz1Xn3Udr/zj1UCXI4QQfunpDJNGkxACAKfTyeO/foo///ePzF/1b1a/8xF5ew4GuiwhhPBJb2SYNJqEEAB8s3kXGVnppA9Kxxxm5pzZ0/hs5dpAlyWEED7pjQyTu+eECEG3Xv4zKivr2m2Pi4vgn2/+tVPHLCsuIzk9ueX3pLQkvtn8badrFEIIb3oiv6B3MkwaTUKEoMrKOobd+ni77bv/eU+nj6m1l3lLQmFgpxAipPREfkHvZJhcnhNCAJCUmkRJQUnL76WFpSQmJwSwIiGE8F1vZJg0moQQAIwYP5zDeUcozC+kqbGJD99dw2nTpwa6LCGE8ElvZJhcnhNCAGAyGbnnd3dy33UP4nK6uHDuBeQMyw50WUII4ZPeyDBpNAkhWkw5ezJTzp4c6DKEEKJTejrDpNEkRAiKi4vwOmgyLi4iANUIIYTvQjm/TthoUkoNB15vtSkX+LXWuv3QdyFEr+jKbbn9ieSXEMEnlPPrhI0mrfUuYAKAUsoIHAEW9XBdQgjRZZJfQoju5O/dc+cC+7TWsraCECLUSH4JIbrE3zFNVwKyimeQ2PzZVlYsXElpQRlJ6YlcMHcGE04fF+iyhAhWkl9BRPJLhCKfe5qUUmHAxcAbHTx+s1LqC6XUF6vfWNNN5YmObP5sK6+/sJD46eFM/M1I4qeH8/oLC9n82dZAlyZE0JH8Ci6SXyJU+XN5bibwlda62NuDWut5WutTtdannvuDad1SnOjYioUryb4kjfjBMRiMBuIHx5B9SRorFq4MdGkihD3yv39hzimXc/2MGwNdSneT/Aoikl+iJ/RGfvnTaLoK6doOGqUFZcRmR7XZFpsdRWlBWYAqEn3BzMvP5y/z/xToMnqC5FcQkfwSPaE38sunRpNSKgKYDrzdo9UInyWlJ1KVZ2uzrSrPRlJ6YoAqEn3B+MnjiI6NDnQZ3UryK/hIfome0Bv55dNAcK11HSArdwaRC+bO4PUXFsIl7jO0qjwbeYsLueLHc73u318GXfaX99mssryKvz/wKPf++QFi42MDXU5QkvwKPv7mF/SP73Z/eI+thWJ+yYzgIar5i7Ri4Up2F+STlJ7IFT+e6/UL1jzoMvuSNLKzR1KVZ3MHVqvj9AX95X22tuqN5Tjyd7Ny4XJ+cMtVgS5HCJ/4k1/QP77b/eE9HisU80saTSFswunjfPoytR50Cbj/vsS9vS99GfvL+2xWWV7FpuWr+Mdlady+dBUz5l4YMmdrQviaX9A/vtv94T22Fqr55e/kliIEdWbQ5ebPtvLIXY/xs8sf5JG7HguJW4H72+DSVW8sZ/YQxdAUK7OHKFYuXB7okoToEf5+tyW/gl+o5pc0mvoBfwddhuocKv1pcGnzWdo1p7jPSq85JYZNy1dRVVHVpeP+9s6Huf3Suzi0P5/Lp1zJstff645yhegSf77bkl/BL5TzSy7P9QP+DroM1W7izgwuDVXNZ2kJUe6vcEKUqeVsrStjA37z1C+7q0Qhuo0/323Jr+AXyvkljaZ+wN9Bl6UFZWRnj2yzLTY7it0F+T1ea1f4+z5D2ZZ1X/FRQQOvbi1os31A2VchM6BSCF/5892W/Ap+oZxf0mjqJ/wZdNncTdx8pgah003sz/sMZX+Y/5dAlyBEr/L1uy35FfxCOb+k0STaCeVu4v42z4kQoi3JL9GTpNEk2gnVbuL+OM+JEKItyS/Rk6TR1M91dGYTit3EoToAVAjROZJfordJo6kf62tnNqE6AFQI4T/JLxEI0mjqx/ramU0oDwANFiUFJTx876OUl1ZgMChmX3URl99waaDLEqIdyS9xrN7IL5ncsh/razPQXjB3BnmLC6nYV43L6aJiXzV5iwu5YO6MQJcWMowmIz/91a28tPo/PLvoKRa9tIS8PQcDXZYQ7Uh+iWP1Rn5JT1M/1tfObEJ1AGgwSUhOICE5AYCIqAiyBg+itKiM7KFZAa5MiLYkv8SxeiO/pNHUT3gbMHm8W3OD5dZXf+sIxQGgnbV+zUbeWvAWhflFpGWmctnVlzFl2qRuO35hfhF7du5l1IQR3XZMITrr2CwYNnoomxZvCur88lb38WqR/Ar+/JJGUz/Q0YDJK348lyt+PLfdmQ0QFAMs+9pAz+60fs1Gnn9uHtlz0hmUM4bKAzU8/9w8gG4Jnrraen5922+589e3Exkd2eXjCdEV3rJg0+JNTJw0kd2r9gRlfnVUt2RYaOeXT40mpVQc8C9gDKCBG7TWn3drJaLHHG/A5INP3tfuy/vIXY8FxQDLvjbQszu9teAtsuekM2BILID77znu7V0NHUeTg1/f+hDnXXIuZ15wRneUG1CSX6GvoyzYvWoPDz55X5t9gyW/jld3f8+wUM4vX3uangBWaK0vV0qFARHdXonoMf7eyurv/h11P3e1i1xuwe1YYX4Rg3LGtNkWlxPNrvyuDXrUWvPoA4+RNSSLK268vEvHCiKSXyHOnyzoTG5IhvWuUM6vEzaalFIxwJnA9Z6iGoHGHqlG9Ah/B0z6s39H3c/7dxxg08ZNXeqW7msDPbtTWmYqlQdqWs7UACoP1JCWmdql4277Yjsr3/6A3BE5/GTmLQDcdP8NTDl7cpeOGyiSX32DP1ngb25IhvW+UM4vX3qacoFS4AWl1HjgS+BurXVtt1UhepS/azFdMHcGLz37MgnTorEkGbGXOjm6poZrb/thu31XLFxJ9Khwdi0+QF1pAxFJVhJHxbNq8Wom3D68S93SobyGVE+77OrL3GMA5rjP0CoP1JC3pICbbrm5S8cdN3EsH+d90E1VBgXJrz7AnyzwJ79AMiwQQjm/fGk0mYCTgTu11huUUk8ADwL/13onpdTNwM0AN/7mes79wbRuLlV0VmduZXU2uCj66Ch2WxOWKDMGu/d/Kvm7j2Co0Ayak0xUVji2g/UcWlJCbWWd1zlU/OmWlltwO9Z83f+tBW+xK/8gaZmp3HTLzd1690kfIfnVB/ibBb7mF0iGBUIo55cvjabDwGGt9QbP72/iDp02tNbzgHkAr+2Yr7utQtEt/LmVdcXClYy8PqdNl3LFvmqvZ1gu5SRzZhIxue5hIjG5EWTMTMD2XF23dEv3p1tw/TVl2qSQCJkAk/zqI3zNAn/yCyTDAiVU8+uEjSatdZFSKl8pNVxrvQs4F9jZ86WJZr0950hpQRn2r6xsemobTfVOzOFGMk9Lpbqgod2+JqOJsGgTjgYnRosRp91JWLQJi8VC3uJC6ZYWASX5FXjBnF8gGSb84+vdc3cCr3juPNkP/LjnShKtBWKeD6fdxeFNxeT+KI2YwRFU76vjwKtFRKrodvtm5KZjrDPgNLuwN9VjMpsx1lnIHpHFBXNnSLe0HzTafUO8CnQlx6E9dYYWya8ACfb8Asmw7hIS+QVdzjCf1p7TWm/WWp+qtR6ntb5Ea13R6VcUfmk9z4fBaCB+cAzZl6SxYuHKHnvNuoZasuemEDs0EmVSxA6NJHtuCnUN7cfOXjB3BofeKcKWX492amz59Rx6p4gL5s5g/44DHNxziPKScg7uOcT+HQd6rOa+wO6yg0sRtG0SDbiUu84QIvkVOMGeXyAZ1l2CPr+gWzJMZgQPcoGY56PJ3kTc0Gi0U6MdLpQyEDc0mr32Aq/7ext0uXb552zdvoWca9OIHRJF1V4b77/2PgCX3jKnx2oPZbVUQx1YDBZUEJ6uaTR2l91dpxA+CIX8Asmw7hDs+QXdk2HSaApygZjnwxpppeZAHfEjvuvOrvi2Bmuktd2+HQ26XP/Ylwy/JbPlGPEjouFKWPXSagmcDmilsVGFLZjP1IIzC0WQCvb8Asmw7hIS+QVdzjBpNAW57lpU9+3nlrBq8WoaahuwRlqZfsm5XHrLHK/HmH7Jue4zqitpOcM68Foh519yfrv983cfIfumCW1eKzY7CqfDReyQY27XHRJFQ21+UC2mKYToOd25KLi3DMsdndOl/Lpg7owOe8M6yrD6mkM8ctdjkl/9lDSaglxH83yA74tSvv3cEt5f9n67bubCg0UcLs73upDv+ZzPqpdW01CbjzXSyvmXnE/u6Jx2r3nwoIP9Kw8zZOaglteryrNhNBmo2mtrc7ZXtdeGOcwsC1gK0U90R36B9wxb8er7GBeZmPDT4Z3Or9dfWIjFYvHaG+Ytw8q/qUaZFfHTwyW/+ilpNIUAb/N8+LMo5arFq8m5Nq1dN/MXz33JlPvGez3GsNFD29WxYuFKUmcMwG6upyCvGpPZzODLMtj18iEShsW1OZM89bRT2PralnZne1FR0bKApRD9SFfzC7xnWNZcF/vnF3Qpv1JnDKB0ebXXqQW8ZtjrhWRPT5f86sek0RSi/Blg2VDb4LWb2elweZ3x9out33Agb3+7nimXDUZOz8QSE0aYJRyn3YmhSWPESMWq+na35b793JJ2Z3trV33e5Vl2hRChzd8B4t4yLDrbisPubHcMf/LL3mSn6mg1P3nweq9TCxybYcZGE8Nn5/hct+h7pNEUovwZYGmNtHq9VGY0Gbwew6EbGX5l+wGQ3zx7kMYaB5EZ4QCYrEZqahwoo+LBJ+9r97qX3jKn3YDJ3Tv2yAKWQvRz/g4Q95ZhNXkNmCzGNvt1Jr8cTkeHs3Yfm2GP3PWY5Fc/J42mEOXPQpAdDYw89bRTyFuc3+4Y2qW990w1uji0uITGcx1Yk800lDRRtLocg/b9n9EFc2fw37+/iMviaHN77/X/c12XPxMhRGjwdyFbbxl2cGExYSYLFfuqJb9Er+mRRpOtykZEdAQGg09zZ4pO8GchyOYzpWMvlbW+e671MZ57+F9ee6YMRgNNtU6K1pTTZHNgjjLRVOskKWGAX7UbrQaSpyW0WYFcCNF/+LuQrbcMu8AzuFvyS/QmpXX3T6rw8L1X6F2l1URGR7Rsq65rYEBuGkq5J0mITohm2KQR3xViUIRZwrq9lv6oq7f0t9ypcmVam56pMJeF5NkxDBgb07JGU/m2aipX2UlKT/Tp9R656zHip4e3mxOlYlW910t8IjRcOfpHfWYGJ1mwN7B6O7/s6xSX33SpT68p+dU3DR0wglPSJvuUYT3S0/TLy09vt83e2ERx+Xct8i0HS/nixVUtv5dX11IfFobR5L5GXWtvImlYeksjK2tsLkkDv7tuLL1Y3nXHWk8d9UytXfU5GePSsVXVtKzRlDQ0gf0LdjDsR5k+vV4gZggWQoSG3s6vjHHpfP3uLp9fU/JL9NqYJkuYmUGp33WDDkodwOzJwzvcv97eSH5xJQBaa95e9SVb6tyrVB+tqYPoKAwGhdYaHWVlQIa7QWUMMzH2jLH9tlHVeq0n6PwtsR0N4m4qc5A6OKVl2961eURnhPv8eoGYIVgIERp6O78q9lXjcDp8fk3JLxG0A8HDLWEMG5Tc8vvPs1I63PdgUTlVtnoACitsvP/kYpQB7I0OqlBYwi1ol8aaFkdMchwAabnpJA9M6tk3EQDddSbkbfZdr4M3lxQy+oohPr+evwNAhRD9R6/n1+JCDNro8zQokl8iaBtN/shq1YM1Djh/ovcerG37Cqi3NwGwcsVG8uoaASipqccS5/7SOA0wcMJgUBCfHE9GbnrPFt/NuuNMqKMZxM/nfK748dw2Ay8TY5OwxlnaPP94r+fvAFAhRP/R2/nV/Luvryn5JXpkIDjrngrZgZRlDAnGsQAAIABJREFUlTZ25hUBsCmvlMPVdQAUVtcTEReFS2sShqYTmxwHSjF4TE7LOKxgsPmzrTz/6L9xWpw4apswRZox2o2MGjuK7V/vaLf2nDe3zbyL+NMise2ro6GsEWtiGFGDI6hYW8stv7yxzYDJYaOHsvaTtSRMi25zN8m1t/1QgqQfkYHgojt0lF83PfAT9u844HX9zGP5k18XzJ0BwEvPviwZ1o8FfCB4KEuMi+LMCe7LTc1/t6a15tMt+6ktLafO7mD1h5sxm01U2urRUeEYlCJhcBqJWcmEWcN6vadq/44DNOkmMs5LIDw5jPqSRvKXlPLlhq8Y9pOBbc68AK+hU1tZh+FbTeacZKIGWbEdaiB/SQm1FXXtBkyufW0tdUcbaPqosc28JUII4S9v+XVk6VGW/nc5+SWH2vUeQfsM8ye/Xn9hIRMnTcTZ4KLoo6OSYeKEfPqXoZTKA2oAJ+DQWp/ak0UFM6UUZ04Y3PL7ZWeObbfPx1v2czSvkIOl1Xy2dAMGg6K0uo6w2Eg0MGTqKCJiIoiKjSIuMbZb61u1eDW516YRlWMlzAgRaVa0U1P0cUW7GXJXvbTaa6PJGGYgY2Yi0TnumXOjc8LJmJnIrn/mtxswmTAtmqaPGjn9V6e0PL9iX7WsxSSChuRX6PCWX5aYML55di8jb8vyKcP8yS8ugVX/WM2E24e3m0ZAMkx4409z+mytdVmPVdKHnDU+t+Xn99bv5PElazhQVE5O6gB+etEZ1OWX4HA62V5Qjk0r7A1N1JmNWMItJAxKYuDoLCKiI4j2jLPyNqixo0trDbUNRGWHY1QagwKjgvCUMJx1x6zRNCSKhlrvgystFgumCCPOBheGMIWrUWOKMIJLtRswaUkyYrc1tT223IIrgo/kVyd0dc4k6Hp+RWWHo53eZ/n2lmH+5FdsdpR7XTtZD1P4SPoge9B763dy/2vvkjsnhdOzkzmaZ+OXby7nz1fOZuaUtj1UjU0OANbtPMSez3ewq6iCeoORDRt3sm3/fhJPiSHntBTsFY2seGcF4P3SmiXCQvVeG0kjIgEwGaC+uBFjxDFrNO21YY20eq07e0QWhhoHdtWIdmmUQWGoMRERE95uwKS91Iklytz22HILrhAhrzvmTOpoUDb4nl8Vu20oo/I6y7c10kr+t4fYtuRzoj15FhMWTvF7FbgsLtCAAoPdgHIqDq4uIHt6Rsv8f1V57mPINALCV742mjSwUimlgee01vN6sKY+4/Ela8idk0KS58uYNDgG5ri3z5wyqs2+YWb3f4pp43OZNv677SlLP2DkjQOJzYmg8ItKDFoRMdDK0vnLMFc2EDMoCaPZxJhp4zGZTWRnDWTfawdwXJxEeEoY9cWNHH63FEeji+KN5VhTwmgobuTw0jIuuOR8r3VfMHcG8595ibjTLUSlWbAdaaDyMzvTLzmXTYs34ZrtxGG2Y2qycnRNDQa7qd36T3ILrggikl+d0B1zJq1avJqca9OIGx6No9FB3PBoco4zNMBbfuW/U0p8dBz7FxQycFZTmww798Jz+HjeUt75v6tb5uabNTSF2194k/AzIt35VWin/tMGHrziHP794QZKNx/FkmHC6DJTvq2Gobm57Hopj8FzM0kYHtttGdZob6SxwXOH9sESDu88CICtwoa10YHRYKDG5SImJa7N86ITYhg6qf0d4OGR4UF101F/5Wuj6TStdYFSKhlYpZT6Vmv9SesdlFI3AzcDPHf/Fdw857RuLjX0HCgq5/Ts5DbbErKj+KzosM/HqKm1M3pkNC40Kd+LxWg0kn7GANbf9y3zb5xBeU0dpRU2lr77Obb6Rgq+PUJDfSP7XytCOzXmWBOOBifaAYUflNNU68AcacKszOSOzunwdeur6mlaU0dRvRNjuBFHrSJ3dA65o3N45YkFVBYdJS41gWvv/iEgt+CKoOZXft34m+s59wfTAlBmcPFnzqSivCL2f7W33XZbeS31hTHUFTSgXS6UwYBSClt5Leve/gyA4VNHkeCZNqa+0oazwcWRlUdx1Dsxhbsvs5nCwazM7TJs+EnDqE5M4PcLPyUpxt07tWPfEUqP1MDianSjCxVmQNkVJw/LZPKobP7nubc4vKeMtMQ4nrp+DueeOoxHX/+YN//zFdtq64mIimDimSdTV1LdUqMvtNZU5BUTbXUvB+asrSc3yT1mNSXKyk8mD0UphcloYICn1uraehoaHW2OsyWvhC+PeV2tNbtKKgmPjmyzvbaugahByZjMctGoKxzjmzjl4sk+7evTJ621LvD8XaKUWgRMAj45Zp95gPsMLoSnHOhOOakDOJpna+lpAjiaZyMn1fcFIqMjLVTtrSUi10KYSdHodFKz3050pAWTyUhyfDTJ8dGMzk0DYOP2nVhnRBARb8HldHFocwUHd5VRtauWgWOS0EoRkRaO0WrkvdffZ/xpY1u6qpstXfAeWRcnMHZcJCaDwuHSbNtay7IF73Hnw7eTYlK8eWsWP11ay+AxuUTHRUkjSQQtf/NLphxw82fOpK/eXc/vZ0zg2Hu2F0RbCQs3YB1kIcwIjU5oOGQnNtrKfaMHUlFTzz9WfcXZ154HQHhKLJN/mNpuUPbmf+xiwq3eB2s/+OR9lJdU4HK6AFi1ah2ZFycyYkQERgM4XbDp/aP8fsH7vPfHW0k3G1hyaya3La1jxqQRJMRG8tdbZ/HXW2cB7oZMZU195z6zM0YS7scaqjGR4cS0bQcxPT6a6ScN9v6EY7hcLg6XVPpTovAiamDHk2cf64SNJqVUJGDQWtd4fp4B/K7z5fUf98yZxv2vvQtz3D1MR/Ns7F9SzJ+vnO3zMe6YeQZ/efUjcq9IYcCQCGr21rH/9WL+d+bZXvdv7t0yGN1d1aOnh+PKhZp9tVQeqqbiSC1RA6xEJ1g5uq2GZX96lYjYKBxmE4k5KeRMGMyRfUfIHR7Dx/MOU3u0icgEM8knxbB/3xHWLvqYi4cYGJpi4eIh9Xz29hpm3jCL6ooaXvj189zw+5tbBrALEWiSX53nz+zXBoOhzSTDze6efRZ/We7Or+ghEdTurePg8lL+d/bZDEodQExkHXxzpGX/jnq3GmobaKi08/lfN2MrrCMqLZLsaemUFrjH9g9Ijv/uGEdKyB0fwxdvFbfklyXezL6CMl5cto5ZQwwMT7Ywa0gD85eu5d5r3HM1lVXauOWRl5n382vbLPkVzAwGQ8jUGtTiYk68j4cvPU0pwCJPb4QJWKC1XtG5yvqX5nFLjy9Zw2dFh8lJHeAZBD7qBM/8zm2XnM4L73xMwX8L2G13EWUxkIiZ279/htf9vfZufVGNCjeQfH4cWenJ1Bc0cmhRCdlp8bz8U/fZVWFZFcXlNbz95ifUl9ex/c064kZGknBWPE67i72fluNq0OxYtZZfXeE+9lWnRHHV62s5/dJprF30MYaCvJZGlBBBQvKrk7pj9mt/8+vY3q2iL8so+rIMo9HE7mV5DPp+EhHpydQVNLLluV1EYWXTk29zxO7gkv91N+Yc9ib2fVZB1veTycy0UpvfwIGFRTRUN/LisrV8cEMCANedHMnchZv40azTSIiN5MVl66goym/TkBLiWCdsNGmt9wPjT7Sf8G7mlFF+NZKO9eKyddwxJYZ7z/xuPqe/fVLV4RfbW+9WyeeVZF6WjCU1DLNZ4UoNI/XsARg+dV+FKKu0ccdjrzLv59fyu6unsXztF9jPUAwYE019oZ36skZqos00FNiJaKjh6dWNbMyr5feXpHHxEAMfvPI+ez7eyB++Z+BX73/G6ZdOa+ltkh4oEUiSX10z4fRxXbr07m9+XTB3Bguefw1LShHK4SI2wUL1zhpiE2KIO9uCNTWMME+GRWZbGdQYy8/nTGHS7U9w7k0XEh0XRUpWCpapmpjsKIwWI8ZsE5kzkyl5t5ohiXDVgkJizDBvbjqzhhiYv3Qt1100lSUfbuSXU408/NHGloYUtO2Bat4m+i8ZPRbk1ny1m4ISOwu2lbTZnl6822voeOvdirVaSc2JpbaynjoNSkHyoBjy64oB2p1hVdc1MHZMDmU1NgwWAwOyoxg2NpUPfrGFOmcYi7fWYtGNXDHvMOYwE3bnB4waoDlSbmBcpLNNb5P0QIlQs/vrPYEuIeSUFh31un3p2u0UlNqYt8nWZntW/s42+eVocrB/+wGKduQzJjaNmtIKiqtsDGiIJmX0CDbs2otJR1K1pZoqz3PMRiPVdQ28uGwd9qoK3n76Tb530WlUlVUzYcwwaqtt2JvqMZnNJAxKZE/xET6uMGHCRVqEk4lPHWZAdDjpxbsBOCujkaxoJ2elN7Zp1EkPlGhNGk1B7p2/3uH3c47t3Trnf5/iaF41I0eEEW5W1Ddpvvm2moGJsZRV2lj68SaevTSR25a6u6pzUgfgKG1i6OCklmOU7qtmyuhsXnngOube/wTPzorgtqV1/PNXN3HTb+fx8NQ6quwwsLKJ/774HraDRzFHR7Bx9VoenRnHo6vWtumBEiJYTSwtD3QJIWfqWWP4alf7u+qsgwZy6e2Xtdu+771NfLUrn2GZyVhKKvj8yUVMH57Bj88eQ+aVbS/d5RdXMOtXz2E7WENmpgmLSWF3aByZYaS4Iln68SbmXZbELz/YzIgxOWREhbPr3/tIOTUJc4yVsq3lKLuLqeOyefPXN7TJrzf+cg9aay772eM8erqTrHgTF+Y4eMDT26S1bpeP0tvUv0mjqR8YMiCeZSu2kRGdhCXTQvURO6UrSpk0ciwvLlvHzFwwNlYzM9fM/KVrWy7xOWdrasLsRDdaOPhuCX++cna7gZQPPP0GZ2U0MiTBRGqUkVPSTRjMEJfiDpakbBOFpY3ENdXyjzseZ+yUsWSMyWbg8IEkpCUE+JMRor2Zk0eeeCfRxn3/XYVxbC7H3j53/u0Xk5KZ3G7/yOgIXi2uZP9773HbmWM47+T263w2y0yJZ3LuQJbt3Ebk6CSSMi2U5tupfb8Up8XExYNheEwD14yzcriglD/dMJv7X3uXMKeT6hIbQycncGhpKQ9ceZ7XgeDg7mXKjjdiNRnIjje29DYBHQ4cF/2T0roH7q6VKQeCysy7n+CLbw7gsBhodGrCjAqT3cX4oYNwNdXz1HQwOepwmCK4cxW88Zd72PjNQe6d9zZHSirISI7nbzdfysQRg5h7/xMsnBtNYpSJMpuDqU8dwmZ3kBplxGAAlwtK61wMz8nE1VTfZt+5C2t44y/3sH7XYTbtL6K00cnR2gbC46PJPmUIaYPTiYmPPvEbEkHnytE/8mmF8JAg+eW3n81fzcS7vu/387TWrFu8loTSKn4z9/R2058085phDS4sZiPv/3hAp/Nr7sIalDmc3QfySYow+Jxh0tvUx6SOg9yzfMow40MPPdT9BeRv7IGDhq6yShvXPvQfzps4kgir73N4NHtv/U7ueOYNfv/K+yzdsJ3EqEiGDkw68RM9SsqrmZpQyZ/PNHDbOBM/GWchKT6OEmck56Q3kBpWx0NrGpiUqgkLC2NrsZMLp47hnRXr+OsZmq8LTdx37UxeXLaOYeYizh0aAUBEmIHSGjtmk4k3fpTBp/vreeP6TMItZoockUwf2Nhm36M1jWwpcvCjmZMZMzCR1xZ9xDO3zOKycdmU7jzIN59uZ++m3Wz+ZBul+aVEJcZgsYa1zPQrgteY5Am/DXQN3Ubyqw1f8mvllgNkdNBDt/mzrfz3sZdY9O93+PqzzURFR5E6yD0vjlKKQSMHsb+4gobDpYzI9J5r3jIszBKGyWhkUoqj0/l1tKaRYkckV4wy8djsJJ8zbOiglC5luggyUSkQn+1ThsnluV7QlYGE3tavu/+1dwF8vitvzVe7+fZADU9/0oBLawxKYbE20eisZocRHq9rICUSLnu9lsgIJyNK3QMjv5dcz5AB7r/nL13rdVB6SUUtTU6Y9PQR4kyNLYMry6rzyS+0djiAvfkzefm9z7n3mhlcNW0cV3n2abA3sb/gKCtXfcnqgnJcBgPmpFgGjc1m4IhMwiPD/foMhRCd50t+ubRGa92up8jX9esmzpzEi39+ndGZiQz1cjnPW4Y5tKLRCZcd0p3OL6Alq55eV+lzhgEyOLyfkkZTD/M20Nqfrl1/1q/ryH/+73ouuffvGO0O5s2O4OZ363BZIvn3Q7dww0PPYQxvu/2xu+dyw2/+yR+narLjTMwe4uAXH6xnyeM/81p7WaWt3eDK473HE30mVouZUTmpjMpJBdxd+GWVNt7ffIAvP9lKrcFIU0QYMUlxjD17PBFRES2TeQohuo+v+XX11JE8+ruXOOe22W0mmjzR+nVaa/Z+tYctyzdyUkocyR1cnveWYXZjNAaDwtxU3aX8an6fvmZYy74yOLxfkv/T9LC2Aw8NLYMLwf3lu+zBf3K0qrbD5x8oKichu+0dZwnZURwoKvf5GC8uW0eSuYE5w80MTzQyZ7iZRLN7EHdH27+XXM/geAMGpRkcb2g5W/P2msd7j/5+Jt4opUiKj+aHZ4/j7zfMYN715/LIeeO5OSuRw699xPt/fp2lf32T9e+u42hROT0yTk+IfsjX/Jo4LIOXbruQ9c8v52jhd9MPlBaUEXtMfsVmR1FyuJT1S9ez4Bf/YcmfXuSpq8/i/644k9go773I3jLMXm8jVtm6nF8nep/+fCai75NGUw9qPku77mT3Wch1J7tvj23d2Gju4u1I8wzfrbVev86XY7y/4Ru2FNQzeaBiZ2kTkwcqthTUs3HHwQ625/Hy5joufKWOyf+q5cJX6nh5cx3L121v95oneo/+fia+ShkQw+jcNB668izm334R/7r2bG7NSqLsnXWs/OOrrPj7W6xbso4je4+c+GBCiHb8za9wSxjzbrqAL/7zPqVHSoHvZvhu5mhwsOOlvag6B1cNiOTM9GhinFW8+v6G49biLcMOVdjZWdzYpfzy5X3685mIvk8Ggvegf761xuvAw+aBhA//622enRPL31cdZNaZp3gdUJgYFcnCt77CkmwmPNZM2YEa9i8p5v8uP5/4qHCfjlFaXs2k+CquODWZpPhoBiXF4MSIMyqFH4w0cebQeG5fdJSfnJ5BuCWMfTUmrhju4g/nWLhvqoUZQ0wYDQaqTAmcN3FUm9csq25gTHip1/f4vXHtF5083mfibX9fmUxGEuOimDYmi0smDWPOhFwGaScHvtzL+mUb2PvVXo4cLCY6MYaIDs5mRefJQPC+pzP5ZTIZufCkXN5+7WPyi6sYMXEka1/bgEM7OLTqCKVflGEoauLZ2y5nTG4af/r3ohPmF3jPsPUH7cwYFs6M0Umdzq9ZZ57S4QBxb5nUU/klAsyPgeAy5UAPuvhnT1NQUtZue3pyItNOHgZHvuTeM2P52ydVkHFKhwMK31u/k8eXrOFAUTk5qQO4Z840Zk4Zxd9eWenTMTqqo6y6gcQYK4VHa0iyOiltMJKWEE1eURVmgybCrIkKU9gaNXVNCovVyl1zz2nzmi9uacRkaP+fOz050evEnMf7TDozkac/dh4o5PX1uzlwtBpDdCTxQ9PJHJFJek5aj75ufyBTDvQ9Xc2vl9ds4+VNu6mvslF0+AhO7WBweqLf+dVRLUfKajAbwemi0/lFximeAeK+ZVIg80v0ID+mHJBGUwA0DyTsyvwf3XEMgF0Hi7nozkd5e24Ely6s471nHiQ+OsLrsf/5q5u49Q/P94l5S7TWbNiRx2d7CtldUkWNwUDO1JEMHJLRZiCr8I00mvoPf7KnsqaOuOiILh3jePprfoluJvM0Bbfu6OI90TF8nRvqpj/O55yUGiZmGHC4YMH6I1RW1zHMXMSoZDM/eCGPOWNjsTc6+PeafR3OWxJqXdNKKQYmxzN1ZCazTx3KnPE56EMl7Fy3k89Xfsn+rQdQ4WYs4WGEWWQelhORy3P9hz/5ZbWYO3UMyS/Rq+TyXHDrji7eEx3jb6+sZOmqj5k1/awOu7ybz9IWz7ViNkCTCy5Z2EBW5kCqqqvbXbZrvpzXlbpDRXVtPau+3senewopqLMTnTqAceedTMqg9nPICOlp6k8kv0SfI5fn+reyShuX/exxfjm5kYc3hvH2X/+HhNhIyipt3PLIy8z7+bUkxEby/fufYbTxINeOM5EVa+BglYuXtjrY4czikZ9ezkV3PsrTF1i4Y4Wd95550Oukc/1FfnEFr67dyYHqeo7amxh+2mgGnzQUa4Ql0KUFBWk0ie7SUX41P9acYTc+/F/JL9E9pNHUv/3tlZVU7vqMa0Y6eeUbI3HDT+fea2a0O3vLnvMg9oYGrwMmTxqeyRCdx40nmfnX103sVdks+vNPA/3WgkKTw8m7G3ax+pt8qpqcRGclM376ycQlxgW6tICRRpPoLh3lV/NjzRn25MIPJb9E9/Cj0eTzPE1KKaNS6mul1NLOVyZ6WlmljSUfbuTCHCdZ8SYuzHGy5KON7D5U0jKzb/O8Il/M/xUjs5LYcM9gttw/jA33DGZkVhKvPXwL277dx6UjTGTFGbh0hIlt3+5jT37JiQvoB8wmI5eeNopnbjyfF2+5gLtHZ1Kw8BM+fHwRS59cxL6v9+JocgS6TNGK5Fdo6Ci/jlbVtpmdfOnHm1j51L2SX6LX+TO55d3ANz1VSF/W0azdvszm7e8xXly2jrMyGsmON2I1GciON3JWeiMPPP0GM3PB2FjNzFzF/KVrW2a2TYxyr6aTGGVi1hADtz36Et8fbiQn3oDVpMiJN/D94Ubuf+qNbvpE+g6DwcCIrBR+d8005v34XJ65bCoZBwpY87c3WfbkIr7d+C2N9sZAlykkvzotGPKrOa9aZ9gDT78h+SV6nU9rzymlBgIXAQ8D9/ZoRX1QRwte+rOQr6/HWLXxG7bvqWHpTgMGA7hcUGRzEmWp4w8/jMfZ1MiFg03cuWoTYdYoyiraL2B5qKiSl0s0K/Y2YVDg0lBWpzGHHereD6YPio0K5/pzJ3D9uROostXz/lf7+Pwf2yixO4kfks5JM04hMkZub+5Nkl9dE+j8Kq1zMaxsO9pRz1PTacmwf63P48DhcBZss7d5Lckv0ZN8XbD3ceB+wPtqiqJDHS146c9Cvs1d1r+cauThjzYe9xjTJ41kekYdV42P4McLDjH/mkHc+lYJo5MUJmcDWfEmDlY2MDM3HGPmSK9h13rSuZZtnonghO9io8KZe+YY5p45Bq01Ow4UMv+FFRypayQmJ5Vx50yQOaF6h+RXJwVDfr2yuY5PyyxMHVDfJsNunBKPMfPUdhkm+SV60gkbTUqpWUCJ1vpLpdS04+x3M3AzwHP3X8HNc07rtiJDWdvFHRtazqg62t7RMc7KaCQr2tnSVd3RMdyz29r565oykqxOTn78EC4NGw+6WLhVEWNVVDdoMDUxonS319dsPsaxPVDpxd73FyemlGJMbjp/yU0H4NuDxSx461M+q7ARPSiZUWeOk+kMeoDkV9cEQ365pwvIZ4cRnl/XeMIMk/wSPemEd88ppf4EXAs4ACsQA7yttf5hh0+Su0+Ajme99Wdm2ubbbx89vZ4JaSY2Fzp4YG04z//65g6PUVZpazNL7it/uoufP/GKzIQbpPbkl/Dmhj1sL64gIiOREaePIWNweqDL8kuw3j0n+dV5wZJfx5vlWzJMdIvuvHtOa/1zrfVArXU2cCXw4XEDR7ToaKB1RwMY5y9d69fA7o6O8eAzb3L1GBPjUs1cPcbErY+82OG+IvCGZibz88tP45WfzuIXk4ai1m5j8R8XsG7xWipKKwNdXkiT/Oq8zuQXtB3c3R35df9Tb3RYi2SY6G1+zdPk6d6+T2s967g7ypkacOKFco/VvBBm67mUZt79BNv35JEU0XZgpMEURlJseLtjxMXEkJd/mE9viCItykihzcmUedXExkQTZja2ez2ZCTd4ffHNIRZ9tZ/9lTYGDMlg4qzJWCPa/7sJBsHa09Sa5Jd/OpNfx87m7R7Y3bX8OuM/tpZZvjt6TSG6RCa3DE1llTYuuffvROpa6lQki/9+r/tMqoNBjd6uz3///mcYazrI7875rsv61x/Wss2RJZO7hbAt+wp5fs02ihuayJ04jKGTRhAVRJclQqHR5DPJr047NsNmnHYSkUe3SX6J4OZHo8nXu+dEL3hx2TqSzA1U1TaRGNl2YKSvgxq/3pXPxsYm/v1128s65rD8Hq1d9Kzxg9N4enAaAB98tZcl/1pOmVKMOns82aOzMYd5XxhViN50bIa99eGXmAxa8kv0GdLTFCSaz9CM9irmzY7g5nfrcFliWfz3e2Wgo/Cqwd7EwrU7WLE1j/DMZE6ZNTlgUxhIT5OQDBMhqyeWURE9q/kMbc5wM8MTjcwZbibR3NArAx39mdlXBA+rxcx150zglbvn8PszR7P/1Q9599HX2fn5TpmFXPS6QGWY5JfoTdJoChLvb/iGLQX1TB6o2FnaxOSBii0F9azc0PMrP7SelVeEHqUU6UmxPHb9dP574wzGHq3k08ffZsW8ZRzZeyTQ5Yl+IlAZJvklepM0mjqhJ85szp88kjvOSOK0UQMZlZPOaaMGcscZScyYPLJHazl2EUw5WwttZpORH5wxhudvvZBHzhuP/aPNvPPIa2xYtp4euRQvQlKwZJjklwg10mjqhJ44s1nz1W4WbLNz6jMlLX8WbLOz5qvdPVpL21l5Zd6TviQ1IYZfzD2DV26dyQVWE6v/tIAVz74rcz+JoMkwyS8RauTuOT/5s+aSPzoz10hXa2l+/sK57iW5rjs5krkLu+89ieCglOL8U4dx/qnDKDpazTNvfMLq0kpGz5zEyEkjAl2e6GXBkmGSXyIUSU+Tn4LpzKartcgsu/1PakIMv79mGq/dMZuc/GKWP/wKH770AXW2+kCXJnpJsGSY5JcIRdJo8kPzmc11J7vPYq47OTJg19G7o5bOXhIUoc9kMnL9eSfx4h2zuXd8FlufX8abD79C/p7LGpmdAAAYT0lEQVTDgS5N9KBgyTDJLxGqZJ4mP/ztlZV+zc59ImWVNm555GXm/fzalu5kb9t6oxYh6hoaeWb5F2w4Usao809l2IQhmMy+XcGXeZpCQ3fmRkdZ5UuGSX6JoCIzgvcMf2fnPpHWgyCbn+9tW2/UIkSENYz/vXQqjU0OlqzbyZtLN5AyfjCnXjiJMGtYoMsT3aA7c6OjrPIlwyS/QpvD4WTP4dJAl+GT+OgIUhNiuu140tMUIGWVNube/wTPzorgtqV1vPGXe9Bat9smAxpFIK3/Np+nVnxJ7IhBTL74e1jCLV73k56m/sVbfiXERna4XYSeJoeTwyWVfL2/hM+/LaSsyYLF6l6oua6hkaicCRhNwd/vUnNkH+G2Qzxw6SlkpyV430l6moJf20GQ382ae+w2OesSgTRlRCZTRmTyxe7D/OPvb2HKTGLqZWdijfDeeBL9g7f8uveaGR1uF8Gv3t5IQWkVb63fz+4iG43KSuSgMcQkTyTn6ikMCwvd73ydrYbfLHmeyLqvufvCsQzPSun0saTRFADebpW97NWNuLRm0dWxLdvk9lkRLE4dNpD/DBvI5n2F/OOpRTgSYznzqrOxRlgDXZroZR3d6j/rjAkyBUAIqaypY822fA4UV7H9iA27IYKojCHknnY7p8QnYDL1nUXAI6KiOf2ae2moq+Xhxf8isnY7t58/mrG5qX4fSxpNAeDtVtmzMhrZVuwkMSqhZVvz7bNypiaCxYTBacwbnMY3ecU8/vQS7HFRnHn1OYEuS/Sijm71f+DpNzqcAkAyLLAcDif2JgeL1+1m88FyqppMNFgTSR1+MimnjWFSbDwGozHQZfY4a0QkZ1x9N/b6Ov62dD7m5au47fzRnJQ6zudjyJimALj4Z09TUFLWZltJRQ1NTshIjG6zPT05sVMTXwrRG/YfKePRdzfy3IvvyZimfsJbfgGUVTeQGNO+51EyrPdprdl2oJh1Ow/z9eFaqh0mrLHJpI09neRBQ4iMiQt0iUHBXl/HlpWvc2ruAG756f/4lGEnbDQppazAJ4AFd8/Um1rr3xz3SRI6QvQvU+8MykaT5JfoD+yNTSzbsIeN+45S7zRSUKtJGDSMgePPIikjC4NBpmQ8ntHpMUwdkthtA8HtwDlaa5tSygx8ppR6T2u9vktVCr/4On+TEKINya8gIPnVvbbsPcLBkmo+3FFMWZ0LFZlA+snTSZmVS3RcAmMCXWAfdsJGk3Z3Rdk8v5o9f+RMrJf5On+TEOI7kl/BQfKrazbsyGP9nhL2FNdS7gwnJmMY8TmTGHzFEEZHRAW6vH7Fp4HgSikj8CUwBHhGa72hR6sSbfTUAptC9AeSX4El+eWf4vJqdhwoYunXBZTWujBHxWPJGEXqmBmMnzU00OX1ez41mrTWTmCCUioOWKSUGqO13t56H6XUzcDNAM/dfwU3zzmt24vtr2TuEyE6T/IrsCS/Oqa1ZuveAtbsKCD/aB2FNo05LoUBwyYy5KrbGWkJ3bmR+iq/phzQWlcqpdYAFwDbj3lsHjAPkIGU3aijOVHkbE0I/0h+9T7Jr7ZKKmoorbDx1uf7OFBWh7bGEp45lrQxZzE8ezjDA12gOKETNpqUUklAkydwwoHzgEd7vDIBdDwnipytCXFikl+B1V/zK7+4AofTxRd7i9l+qJyGRgfFDSZ0xABiBg4l85xZTE3yf2JFEXi+9DSlAfM94wIMwEKt9dKeLUs0k4UthegSya8A6sv5VVxezcfbDrOvuBq0Zk9JPdaoGBwOB47oDMJj4ohOnMKgSycBMCSElyER35HJLYUQXRek8zR1iuSX6ICtzs6idbtYvb0IkoaSNvwkUoe4Z5O2hkf0i1m1+6LunqdJCCHacDicNDmdAOzKP8qEqQEuSIge0uRw8uHmA7yx4RB1pnhyJl/ApJvHYpaeo35JGk1CiDZa9z5v/PYIu4+Uo7Xmy7wKMIUDUGprxBqfBkBY9AD+GZBKhegZWmuWb9zDyq0FlDSYSR17Gqf+5KcYTfK/zP5O/gUI0Q/V2xtZvmEPLg1Ol4vPdpdhCrOiNRRWN2GNdS8cHZ2STeqI6QAMmZyMVSbSE/3Anf9cjWXsLIZccRNj5N+8aEUaTUL0Mdv2F3CwuBqAb49UcqTGBUBlTT3OiEQMBoVLK9JPmUGY1d1zNGpSLmbPnDCyBIPo74zWSEZMOTfQZYggJI0mIUJEg72Jz7cfANzjLN7bVozBaEZrTWGNE2u0e+XysPgMBgx2B37kqbGMGpgTsJqFCEZVtnqeW7GF4hqn18eLql1y8iC8kkaTEAHmcrn44tt8nC4XWmuWfXUYu8t9I8fR6noM0cmAu6E0YOzZmMxhYITBV47D4ukpGhmw6oUIDSUVNSzfuI9PdpXiCE9i2Hk3MCLT+wnFiF6uTYQOaTQJ0UPq7Y18e7C45feVW45w1NYIQHF1I+YY97ghR5ODyJyTsUTFAjDwoquJjkvo/YKF6GPq7Y18suUAizYexGZJIWvKRXzv3AmBLkuEMGk0CeGng0Xl1DW4Gz95JdV8tL0ApZTnMpmL8Bj3ZbKGRidxQydiMLjnbkmZPJOh6VkAyLKbQvSM5vXc/rFiO3XmeBJGn85JN9+DwWAIdGmiD5BGkxBAaUUNtZ6GkMuleXPdHmrt7vEOxZV16IgBLQ2jpvAkIpPSATBbsxhx7T0o5b6cJo0hIQLjSGkl81ZsZV8lRGZPYPR1DxMeGR3oskQfI40m0SdprSmttOFyueccKq+u5Z0vDrY8XlxRS4MxGmUwoLWmzhBFVFJGy+MDp9xMSkIKAKkGAyaTuXffgBDihEoravjPqm3sLLaj4gcx7IxbOUtufBA9SBpNImRU19a3NIK0hnc37KGyrgkAh9PJjsJ6wiMiALA3NuGIzsAS7llJ3Whl2Nn3YDS6Gz+DTSYs4RG9/yaEEF1SU9vAm2t3sW7PUerDBjBm5u2clpLe0tsrRE+SRpMIiCaHE6fT1fL7l3uL2HmorOX3gnIbpXYzSrnHIdgbG6m3JGD1NIK01qSOPJ+41EEtz5kUnyDjFoToYxrsTdibHLy1dhdbDlZQSjy537uQk88Y7b6TVIheJI0m0a1cLhdvffYN9sbv5j/JK62hpN5I84mg1lBS6yIiPqlln8ikgaSPmdXye0RYGOMTknutbiH8UVtvZ/G6XRy73vnuwioq6hxMyh3gtedjfG4KY3NTeqnK0PPBV/soKq+lsKKWQ9Wa+oYm7OGJmMxhZE68lMFTBjJGximJAJJGk/Aqr6icL3YXttm2q6CK0vq2/yMoqW7AGJXYZlv6hLOIzPiuwWMZFc7Y5PSeK1aIbnTPC5+3+b2guglLVHzbnZQia8pswizhbTZHj40mKTySA6VFXo/92dZPqf1oW7vtTqeD8MZKoiOt7R9rsnPG8EQMXhphg9PiOGloRrvtwWzfkTK+3leMw+levsdotlBY48ISFUti7hjic0ZiHh7GuLTMQJcqRDtKH3uq1B3WPdUDBxW+KqmoYdOuI+2255XUsPdoI4rvwtfe1ES5w9oySWIzQ3gMGeOntTlbjoiOJSH1/9u79xg7yvOO47/nnL3g2Gtce31be4NxuBOlIUEuCqrVpkkbiEsTpS2xYlLahFVIHYFASnOpGtESNc0fFhfTRuZSixKTKCKVkEPaIpUKITUmNZe02EWmxrDG9h7vLvZ6ifdyznn6x1mk7M7M7hzvnJ1zZr4fybL3PbPvvrOyfnrmnXfeaa2AxsK4ZfPGzCwoefDZw02VX+NjZ1Xqfz30s5OH9uvs4NFAe7XqKrwzqOVLFwU+m5yc0O9cvlLt7cXAZ2uXL9aHLkmmWPnZgTc0NHJW+14b1OhkQQMj42rvWqHie35NPR/4TZmZ1lzwPrV3dCby84BzdWXPUn3kou5YGcZMU8r6B97W/701GPn5MwcGNDoZ/tmJkUl1dC0PtHuxU2t+/bcDtwc6L3uPrth46bzGC2BhdZ63SL0XXxH6WVT7bCYnxrXv0IHQz069elBjz70c+ln1zEl1Lw3OhJXLZV2wvFNHT1d0+p0xlRfV1hYu6rlE5699n5Z/fJ16V6xi13pkwpxFk5n1SnpU0hpJVUm73P3eRg+s0d4e+aVefXNg7gND/PzwoN4YGpvzuEqlqoGzhVn3CvHOLi2/5JrIz1d/7EKt7w5fA8FW/8Dssppf89He0amNV14V/mFU+ywq5bKOHTmkyy+iLEL2xZlpKku6091fMLMuSfvN7Gl3D79UkXTLP76Y2AAbZbxS0IrLrpHO4THVZZf16NKYMzb1XwcCSFDd+YX6FNva1EvBhJyYs2hy9+OSjk/9+4yZHZS0TlJk6Gy66S8TGyAAnKtzyS8AiFLXmiYz2yDpKkn7GjEYZN/fbt+q0dEzgfYlS7r09Z2PN23faH3kF+aL/ELsosnMlkh6QtLt7j4S8nmfpD5J2nbn3dp8w9bEBonsGB09o41fvD/QfvihrzR132ht5BeSQH4hVtFkZu2qBc733f3HYce4+y5Ju6Tme2QXQH6RXwCSMuc7J6z23PrDkg66+47GDwkAkkF+AUhSnBd1XSvpJkkfNbOXpv5c3+BxAUASyC8AiYnz9NxzkjKz2y8WRtSixuHScXUeORRoPz0UvcFnXKeHBvVWSN/DpeP65s1bAu0ssMw+8gvnKizDhksn1PbaARXb2qe1k1/5wY7gaIioRY1Df/NpDe0N3iXxanneP9Or5fC+KxUWWAKoS1iGjTywXaUf3RV4EwP5lR8UTVhQhfYOfWD7PwTakwiAZSvXhIbL/u/cOO++AWDlJ2/T0N4dgQwjv/KDogkBUbfWzgyfVNfyldPaTg8NyqtlLVu5Zlr7qYi3vFcnx/XCPV8MtJdHTs57CvrUyRP6xc5bQ37mRKzvB9D66smv2drDMqy09x5VzgwFMoz8yg+KJgRE3Vrb/50bA+1vHTmkob07Au2RV0eFNvX82c5Ac//92+Y9Be1WUM/N9wTaX7/vc7H7ANDa6smvudpn8vK4Vt14tzq63zutnfzKD4omxFaZCJ8lqrzzduw+vDKpow98PuyT0EWQYQssb/29D8qLxWAP5bLOlt4M6Tr+tjtJ7crL7r5Ac4nMr4nx2H2URwZV+mHYa8Li55cUkWEu9e++Q93X3z6jvb5tw5LIHvIrGkUTYrNicJaoOjmuY7tuCRxbsELoVZYV27Xu1t2B9qMPfF6dM67eJKnq1UCbF4vq3f5YoL3//m1qW7Y6bOShY1mypCvQltSuvOzuCzSXsPySarkRJizDrNg27/ySwjNsYvBNDT753ZAMi59fUjLZQ35Fo2hCQ5y/olvf3r030P6l6z4sK8TZHuzctHd0BtoKxWLoWAAgSliGNTq/pGCGkV/NhaIpx6KmYAeP9etUyDS2V4KP1ZZ+9C15tRpYwDj+dvjeIl4pqzoZPh0+MT427evy6ZKqlYq+vGXTjE4W/i0Xp06eYK8UoIkkkV/vtoctwI7KsBN7vqZVf3RXoH1mfklStVIO5pe04Bn2ykN3amywFDgf8qt+FE05FrmX0rc/o+4bvhpoH3j8G5oYnL5mqDo2qtV//Ncqtk/f7O34nm9E9P2HsmJ7oN0rZQ3s+YtpbZXRYbWdv1q9N0/fu+T1e6MWRrqO7b490GoRU+T1cCswXQ00kSTyS6rdclux5Y5Ae1iGtR06oKGf3hfIsLD8qrVXdMFtewLtoRlmpso7pwIZlkR+VcZ+qTWfvVvrNlw8rZ38qh9FUwpad5GdB54aqY7VzqNSnn4V57NcSUVNb1dGh6f3Ua1KqueKzEL3gHrx77YySwQkJEv59W77zPySZskwC8+wmflVr7alq2WFQiDDyK/mQtGUguZfZGcR4WLq3zlj0aRLbct7VGg/L2bfHvr0nBWL6ul7cFpb+fSABp/8brCHSjk4DkmqTIb+Dr1ajv37XrKkK7S9YPWtY4jqJ2rxJtAqMpVf7x6/6sJYPbtc5ZHBQIaF5ZckvfW9L4T3E5ZhXnu6eObvsZ78ksKzZ3J0WMWQJ46jkF/RKJpQl+89tX/a11+6/sN1FEySzLThK/8UaH793s+p0B5cxB2m2NGpv9/7fOwfGXaVFiXqyq2ePmbrB0B6ZuaXVMuwuMwKalvard4/vXdaez35JdWXYUlkzzdv3qI1vRvn1QdqKJowP966O9aeHhpk2hvIuxbNMPIrHRRNORY1BeuVso49sj34DSEvpfRqWQN7gosuw6aZJckqFb2xM2xzS4U8VVdbGDnz+LZCMi+tr3p13rftmK4G0pFEfkn1Zdhw6bg8IsPCnwr20GOTyDDyKx0UTTkWdTWy/fev0apPBZ8EKf3wrwJtVmzT2j8Jbv3fv3NbXXuLfHnLJnV0Tr/N17HqvSoUi3XdimsUrtyA5pJEfknJZFhYfklSodhGfmUMRVMKmr3q92pZQ3t3hLYHVCrhV3WVSl0/s61gDbsii/p9J/EoL5A3mcovKZEMI7/yY86iycwekbRFUsnd39/4IWVfs1f9y1auiT3tu2Lt+kSepLnvyX11HV+PpBZ3ozWRYcnKUn5JyWQY+ZUfcWaadkvaKenRxg4FmK5194NBk9ktMgwpIMOyZ86iyd2fNbMNjR8KWlEjp+obvR9Ms99mQDLIMMymVTOM/EoHa5owL618tdTKYweQjFbNgVYdd6tLrGgysz5JfZK07c67tfmGrUl1jQXGFQzyhvzKDvILjZRY0eTuuyTtkqQHnz288K+hR2K4gkHekF/ZQX6hkep7oRYAAEBOxdly4HFJvyWp28yOSvqWuz/c6IEBTLMjCWQY0kKGZY+5Jz8TzfQ2kC+3bN6YzLttmgD5BeTLlT1L9ZGLumNlGLfnAAAAYqBoAgAAiIGiCQAAIAaKJgAAgBgomgAAAGKgaAIAAIiBogkAACAGiiYAAIAYKJoAAABioGgCAACIgaIJAAAgBoomAACAGCiaAAAAYqBoAgAAiIGiCQAAIAaKJgAAgBhiFU1m9gkze9XMXjOzrzV6UACQFPILQFLmLJrMrCjpAUnXSbpC0lYzu6LRAwOA+SK/ACQpzkzTJkmvufthd5+Q9ANJf9DYYQFAIsgvAIlpi3HMOkn9v/L1UUm/Mds3dHd1zGdMAJAU8gvArBZ3ximFauIcaSFtHjjIrE9S39SXj7n7TbFH0YLMrM/dd6U9jkbjPLMjD+cYgvyKkIf/D3k4R4nzXEhxbs8dldT7K1+vl3Rs5kHuvsvdr3b3qyVdntD4mlnf3IdkAueZHXk4x5nIr2h5+P+Qh3OUOM8FE6do+rmki83sQjPrkPRZSU82dlgAkAjyC0Bi5rw95+5lM9su6V8lFSU94u6vNHxkADBP5BeAJMVa/eTuT0l6qo5+M39vVfk4R4nzzJI8nGMA+RUpD+eZh3OUOM8FY+6BNZEAAACYgdeoAAAAxJBo0WRmj5hZycz+J8l+m4mZ9ZrZM2Z20MxeMbPb0h5T0szsPDN73sxenjrHu9IeUyOZWdHMXjSzvWmPpVHM7IiZ/beZvWRm/5X2eJoR+ZUdecow8muBx5Lk7Tkz2yxpVNKj7v7+xDpuIma2VtJad3/BzLok7Zf0KXc/kPLQEmNmJmmxu4+aWbuk5yTd5u4/S3loDWFmd0i6WtJSd9+S9ngawcyOSLra3QfTHkuzIr+yI08ZRn4trERnmtz9WUnDSfbZbNz9uLu/MPXvM5IOqrbrcGZ4zejUl+1TfzK5+M3M1kv6pKSH0h4L0kV+ZUdeMoz8WnisaZoHM9sg6SpJ+9IdSfKmpnxfklSS9LS7Z+4cp9wj6auSqmkPpMFc0r+Z2f6p3a+Rc1nOLyk3GUZ+LTCKpnNkZkskPSHpdncfSXs8SXP3irt/ULUdlDeZWeZuV5jZFkkld9+f9lgWwLXu/iFJ10n686lbUciprOeXlP0MI7/SQdF0DqbukT8h6fvu/uO0x9NI7n5K0n9I+kTKQ2mEayXdMHW//AeSPmpmj6U7pMZw92NTf5ck/bOkTemOCGnJU35Jmc4w8isFFE11mlpg+LCkg+6+I+3xNIKZrTSzZVP/XiTpY5L+N91RJc/dv+7u6919g2qv1/h3d9+W8rASZ2aLpxb9yswWS/pdSZl9QgzR8pBfUj4yjPxKR9JbDjwu6T8lXWpmR83sC0n23ySulXSTalX9S1N/rk97UAlbK+kZM/uFau/uetrdM/s4aw6slvScmb0s6XlJP3H3f0l5TE2H/MoUMiw7miq/2BEcAAAgBm7PAQAAxEDRBAAAEANFEwAAQAwUTQAAADFQNAEAAMRA0QQAABADRRMAAEAMFE0AAAAx/D9jO+d5lcQUigAAAABJRU5ErkJggg==\n", + "image/png": "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\n", "text/plain": [ "
" ] @@ -231,9 +231,9 @@ "3-fold cross validation:\n", "\n", "Accuracy: 0.91 (+/- 0.01) [KNN]\n", - "Accuracy: 0.93 (+/- 0.05) [Random Forest]\n", - "Accuracy: 0.92 (+/- 0.03) [Naive Bayes]\n", - "Accuracy: 0.94 (+/- 0.03) [StackingClassifier]\n" + "Accuracy: 0.95 (+/- 0.01) [Random Forest]\n", + "Accuracy: 0.91 (+/- 0.02) [Naive Bayes]\n", + "Accuracy: 0.92 (+/- 0.02) [StackingClassifier]\n" ] } ], @@ -284,16 +284,16 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.927 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.03 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.933 +/- 0.03 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.927 +/- 0.03 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.947 +/- 0.02 {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.947 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.947 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", "0.933 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", "0.940 +/- 0.02 {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Best parameters: {'kneighborsclassifier__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Accuracy: 0.94\n" + "Best parameters: {'kneighborsclassifier__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "Accuracy: 0.95\n" ] } ], @@ -352,24 +352,24 @@ "name": "stdout", "output_type": "stream", "text": [ - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.907 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.927 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", - "0.927 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", - "0.920 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", - "0.667 +/- 0.00 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.933 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.927 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.947 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.947 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.933 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.947 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", + "0.933 +/- 0.03 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 1, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 10}\n", + "0.953 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 0.1, 'randomforestclassifier__n_estimators': 50}\n", "0.933 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 10}\n", "0.940 +/- 0.02 {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Best parameters: {'kneighborsclassifier-1__n_neighbors': 5, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", - "Accuracy: 0.94\n" + "Best parameters: {'kneighborsclassifier-1__n_neighbors': 1, 'kneighborsclassifier-2__n_neighbors': 5, 'meta_classifier__C': 10.0, 'randomforestclassifier__n_estimators': 50}\n", + "Accuracy: 0.95\n" ] } ], @@ -445,18 +445,38 @@ "data": { "text/plain": [ "StackingClassifier(average_probas=False,\n", - " classifiers=[Pipeline(memory=None,\n", - " steps=[('columnselector', ColumnSelector(cols=(0, 2), drop_axis=False)), ('logisticregression', LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n", - " intercept_scaling=1, max_iter=100, multi_class='warn',\n", - " n_jobs=None,...nalty='l2', random_state=None, solver='warn',\n", - " tol=0.0001, verbose=0, warm_start=False))])],\n", - " drop_last_proba=False,\n", - " meta_classifier=LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n", - " intercept_scaling=1, max_iter=100, multi_class='warn',\n", - " n_jobs=None, penalty='l2', random_state=None, solver='warn',\n", - " tol=0.0001, verbose=0, warm_start=False),\n", - " store_train_meta_features=False, use_clones=True,\n", - " use_features_in_secondary=False, use_probas=False, verbose=0)" + " classifiers=[Pipeline(memory=None,\n", + " steps=[('columnselector',\n", + " ColumnSelector(cols=(0, 2),\n", + " drop_axis=False)),\n", + " ('logisticregression',\n", + " LogisticRegression(C=1.0,\n", + " class_weight=None,\n", + " dual=False,\n", + " fit_intercept=True,\n", + " intercept_scaling=1,\n", + " l1_ratio=None,\n", + " max_iter=100,\n", + " multi_class='auto',\n", + " n_jobs=None,\n", + " penalty='l2',\n", + " random_state=None,\n", + " sol...\n", + " meta_classifier=LogisticRegression(C=1.0, class_weight=None,\n", + " dual=False,\n", + " fit_intercept=True,\n", + " intercept_scaling=1,\n", + " l1_ratio=None,\n", + " max_iter=100,\n", + " multi_class='auto',\n", + " n_jobs=None, penalty='l2',\n", + " random_state=None,\n", + " solver='lbfgs',\n", + " tol=0.0001, verbose=0,\n", + " warm_start=False),\n", + " store_train_meta_features=False, use_clones=True,\n", + " use_features_in_secondary=False, use_probas=False,\n", + " verbose=0)" ] }, "execution_count": 8, @@ -490,13 +510,185 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# API" + "## Example 5 -- ROC Curve with `decision_function`" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Like other scikit-learn classifiers, the `StackingCVClassifier` has an `decision_function` method that can be used for plotting ROC curves. Note that the `decision_function` expects and requires the meta-classifier to implement a `decision_function`." ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, + "outputs": [], + "source": [ + "from sklearn import model_selection\n", + "from sklearn.linear_model import LogisticRegression\n", + "from sklearn.neighbors import KNeighborsClassifier\n", + "from sklearn.svm import SVC\n", + "from sklearn.ensemble import RandomForestClassifier\n", + "from mlxtend.classifier import StackingCVClassifier\n", + "from sklearn.metrics import roc_curve, auc\n", + "import numpy as np\n", + "from sklearn.model_selection import train_test_split\n", + "from sklearn import datasets\n", + "from sklearn.preprocessing import label_binarize\n", + "from sklearn.multiclass import OneVsRestClassifier\n", + "\n", + "\n", + "iris = datasets.load_iris()\n", + "X, y = iris.data[:, [0, 1]], iris.target\n", + "\n", + "\n", + "# Binarize the output\n", + "y = label_binarize(y, classes=[0, 1, 2])\n", + "n_classes = y.shape[1]\n", + "\n", + "\n", + "\n", + "RANDOM_SEED = 42\n", + "\n", + "\n", + "X_train, X_test, y_train, y_test = train_test_split(\n", + " X, y, test_size=0.33, random_state=RANDOM_SEED)\n", + "\n", + "clf1 = LogisticRegression()\n", + "clf2 = RandomForestClassifier(random_state=RANDOM_SEED)\n", + "clf3 = SVC(random_state=RANDOM_SEED)\n", + "lr = LogisticRegression()\n", + "\n", + "\n", + "sclf = StackingClassifier(classifiers=[clf1, clf2, clf3],\n", + " meta_classifier=lr)\n", + "\n", + "\n", + "# Learn to predict each class against the other\n", + "classifier = OneVsRestClassifier(sclf)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "**Using `predict_proba()`**" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "y_score = classifier.fit(X_train, y_train).predict_proba(X_test)\n", + "\n", + "# Compute ROC curve and ROC area for each class\n", + "fpr = dict()\n", + "tpr = dict()\n", + "roc_auc = dict()\n", + "for i in range(n_classes):\n", + " fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_score[:, i])\n", + " roc_auc[i] = auc(fpr[i], tpr[i])\n", + "\n", + "# Compute micro-average ROC curve and ROC area\n", + "fpr[\"micro\"], tpr[\"micro\"], _ = roc_curve(y_test.ravel(), y_score.ravel())\n", + "roc_auc[\"micro\"] = auc(fpr[\"micro\"], tpr[\"micro\"])\n", + "\n", + "plt.figure()\n", + "lw = 2\n", + "plt.plot(fpr[2], tpr[2], color='darkorange',\n", + " lw=lw, label='ROC curve (area = %0.2f)' % roc_auc[2])\n", + "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n", + "plt.xlim([0.0, 1.0])\n", + "plt.ylim([0.0, 1.05])\n", + "plt.xlabel('False Positive Rate')\n", + "plt.ylabel('True Positive Rate')\n", + "plt.title('Receiver operating characteristic example')\n", + "plt.legend(loc=\"lower right\")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "**Using `decision_function()`**" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "y_score = classifier.fit(X_train, y_train).decision_function(X_test)\n", + "\n", + "# Compute ROC curve and ROC area for each class\n", + "fpr = dict()\n", + "tpr = dict()\n", + "roc_auc = dict()\n", + "for i in range(n_classes):\n", + " fpr[i], tpr[i], _ = roc_curve(y_test[:, i], y_score[:, i])\n", + " roc_auc[i] = auc(fpr[i], tpr[i])\n", + "\n", + "# Compute micro-average ROC curve and ROC area\n", + "fpr[\"micro\"], tpr[\"micro\"], _ = roc_curve(y_test.ravel(), y_score.ravel())\n", + "roc_auc[\"micro\"] = auc(fpr[\"micro\"], tpr[\"micro\"])\n", + "\n", + "plt.figure()\n", + "lw = 2\n", + "plt.plot(fpr[2], tpr[2], color='darkorange',\n", + " lw=lw, label='ROC curve (area = %0.2f)' % roc_auc[2])\n", + "plt.plot([0, 1], [0, 1], color='navy', lw=lw, linestyle='--')\n", + "plt.xlim([0.0, 1.0])\n", + "plt.ylim([0.0, 1.05])\n", + "plt.xlabel('False Positive Rate')\n", + "plt.ylabel('True Positive Rate')\n", + "plt.title('Receiver operating characteristic example')\n", + "plt.legend(loc=\"lower right\")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# API" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, "outputs": [ { "name": "stdout", @@ -602,6 +794,27 @@ "\n", "
\n", "\n", + "*decision_function(X)*\n", + "\n", + "Predict class confidence scores for X.\n", + "\n", + "**Parameters**\n", + "\n", + "- `X` : {array-like, sparse matrix}, shape = [n_samples, n_features]\n", + "\n", + " Training vectors, where n_samples is the number of samples and\n", + " n_features is the number of features.\n", + "\n", + "**Returns**\n", + "\n", + "- `scores` : shape=(n_samples,) if n_classes == 2 else (n_samples, n_classes).\n", + "\n", + " Confidence scores per (sample, class) combination. In the binary\n", + " case, confidence score for self.classes_[1] where >0 means this\n", + " class would be predicted.\n", + "\n", + "
\n", + "\n", "*fit(X, y, sample_weight=None)*\n", "\n", "Fit ensemble classifers and the meta-classifier.\n", @@ -649,6 +862,11 @@ "\n", " Target values.\n", "\n", + "\n", + "- `**fit_params` : dict\n", + "\n", + " Additional fit parameters.\n", + "\n", "**Returns**\n", "\n", "- `X_new` : numpy array of shape [n_samples, n_features_new]\n", @@ -669,14 +887,14 @@ "\n", "**Parameters**\n", "\n", - "- `X` : {array-like, sparse matrix}, shape = [n_samples, n_features]\n", + "- `X` : numpy array, shape = [n_samples, n_features]\n", "\n", " Training vectors, where n_samples is the number of samples and\n", " n_features is the number of features.\n", "\n", "**Returns**\n", "\n", - "- `labels` : array-like, shape = [n_samples] or [n_samples, n_outputs]\n", + "- `labels` : array-like, shape = [n_samples]\n", "\n", " Predicted class labels.\n", "\n", @@ -722,7 +940,7 @@ "\n", "*score(X, y, sample_weight=None)*\n", "\n", - "Returns the mean accuracy on the given test data and labels.\n", + "Return the mean accuracy on the given test data and labels.\n", "\n", "In multi-label classification, this is the subset accuracy\n", "which is a harsh metric since you require for each sample that\n", @@ -730,17 +948,17 @@ "\n", "**Parameters**\n", "\n", - "- `X` : array-like, shape = (n_samples, n_features)\n", + "- `X` : array-like of shape (n_samples, n_features)\n", "\n", " Test samples.\n", "\n", "\n", - "- `y` : array-like, shape = (n_samples) or (n_samples, n_outputs)\n", + "- `y` : array-like of shape (n_samples,) or (n_samples, n_outputs)\n", "\n", " True labels for X.\n", "\n", "\n", - "- `sample_weight` : array-like, shape = [n_samples], optional\n", + "- `sample_weight` : array-like of shape (n_samples,), default=None\n", "\n", " Sample weights.\n", "\n", @@ -813,5 +1031,5 @@ } }, "nbformat": 4, - "nbformat_minor": 1 + "nbformat_minor": 4 }