Pushing the docs to dev/ for branch: master, commit fc7d6e698668b983cee2867b1bf3c65f1384e4cf

Author: sklearn-ci

dev/_downloads/auto_examples_jupyter.zip

@@ -15,7 +15,7 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
-        "\n# Plot classification probability\n\n\nPlot the classification probability for different classifiers. We use a 3\nclass dataset, and we classify it with a Support Vector classifier, L1\nand L2 penalized logistic regression with either a One-Vs-Rest or multinomial\nsetting, and Gaussian process classification.\n\nThe logistic regression is not a multiclass classifier out of the box. As\na result it can identify only the first class.\n\n"
+        "\n# Plot classification probability\n\n\nPlot the classification probability for different classifiers. We use a 3 class\ndataset, and we classify it with a Support Vector classifier, L1 and L2\npenalized logistic regression with either a One-Vs-Rest or multinomial setting,\nand Gaussian process classification.\n\nLinear SVC is not a probabilistic classifier by default but it has a built-in\ncalibration option enabled in this example (`probability=True`).\n\nThe logistic regression with One-Vs-Rest is not a multiclass classifier out of\nthe box. As a result it has more trouble in separating class 2 and 3 than the\nother estimators.\n\n"
       ]
     },
     {
@@ -26,7 +26,7 @@
       },
       "outputs": [],
       "source": [
-        "print(__doc__)\n\n# Author: Alexandre Gramfort <[email protected]>\n# License: BSD 3 clause\n\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.svm import SVC\nfrom sklearn.gaussian_process import GaussianProcessClassifier\nfrom sklearn.gaussian_process.kernels import RBF\nfrom sklearn import datasets\n\niris = datasets.load_iris()\nX = iris.data[:, 0:2]  # we only take the first two features for visualization\ny = iris.target\n\nn_features = X.shape[1]\n\nC = 1.0\nkernel = 1.0 * RBF([1.0, 1.0])  # for GPC\n\n# Create different classifiers. The logistic regression cannot do\n# multiclass out of the box.\nclassifiers = {'L1 logistic': LogisticRegression(C=C, penalty='l1'),\n               'L2 logistic (OvR)': LogisticRegression(C=C, penalty='l2'),\n               'Linear SVC': SVC(kernel='linear', C=C, probability=True,\n                                 random_state=0),\n               'L2 logistic (Multinomial)': LogisticRegression(\n                C=C, solver='lbfgs', multi_class='multinomial'),\n               'GPC': GaussianProcessClassifier(kernel)\n               }\n\nn_classifiers = len(classifiers)\n\nplt.figure(figsize=(3 * 2, n_classifiers * 2))\nplt.subplots_adjust(bottom=.2, top=.95)\n\nxx = np.linspace(3, 9, 100)\nyy = np.linspace(1, 5, 100).T\nxx, yy = np.meshgrid(xx, yy)\nXfull = np.c_[xx.ravel(), yy.ravel()]\n\nfor index, (name, classifier) in enumerate(classifiers.items()):\n    classifier.fit(X, y)\n\n    y_pred = classifier.predict(X)\n    classif_rate = np.mean(y_pred.ravel() == y.ravel()) * 100\n    print(\"classif_rate for %s : %f \" % (name, classif_rate))\n\n    # View probabilities=\n    probas = classifier.predict_proba(Xfull)\n    n_classes = np.unique(y_pred).size\n    for k in range(n_classes):\n        plt.subplot(n_classifiers, n_classes, index * n_classes + k + 1)\n        plt.title(\"Class %d\" % k)\n        if k == 0:\n            plt.ylabel(name)\n        imshow_handle = plt.imshow(probas[:, k].reshape((100, 100)),\n                                   extent=(3, 9, 1, 5), origin='lower')\n        plt.xticks(())\n        plt.yticks(())\n        idx = (y_pred == k)\n        if idx.any():\n            plt.scatter(X[idx, 0], X[idx, 1], marker='o', c='w', edgecolor='k')\n\nax = plt.axes([0.15, 0.04, 0.7, 0.05])\nplt.title(\"Probability\")\nplt.colorbar(imshow_handle, cax=ax, orientation='horizontal')\n\nplt.show()"
+        "print(__doc__)\n\n# Author: Alexandre Gramfort <[email protected]>\n# License: BSD 3 clause\n\nimport matplotlib.pyplot as plt\nimport numpy as np\n\nfrom sklearn.metrics import accuracy_score\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.svm import SVC\nfrom sklearn.gaussian_process import GaussianProcessClassifier\nfrom sklearn.gaussian_process.kernels import RBF\nfrom sklearn import datasets\n\niris = datasets.load_iris()\nX = iris.data[:, 0:2]  # we only take the first two features for visualization\ny = iris.target\n\nn_features = X.shape[1]\n\nC = 10\nkernel = 1.0 * RBF([1.0, 1.0])  # for GPC\n\n# Create different classifiers.\nclassifiers = {\n    'L1 logistic': LogisticRegression(C=C, penalty='l1',\n                                      solver='saga',\n                                      multi_class='multinomial',\n                                      max_iter=10000),\n    'L2 logistic (Multinomial)': LogisticRegression(C=C, penalty='l2',\n                                                    solver='saga',\n                                                    multi_class='multinomial',\n                                                    max_iter=10000),\n    'L2 logistic (OvR)': LogisticRegression(C=C, penalty='l2',\n                                            solver='saga',\n                                            multi_class='ovr',\n                                            max_iter=10000),\n    'Linear SVC': SVC(kernel='linear', C=C, probability=True,\n                      random_state=0),\n    'GPC': GaussianProcessClassifier(kernel)\n}\n\nn_classifiers = len(classifiers)\n\nplt.figure(figsize=(3 * 2, n_classifiers * 2))\nplt.subplots_adjust(bottom=.2, top=.95)\n\nxx = np.linspace(3, 9, 100)\nyy = np.linspace(1, 5, 100).T\nxx, yy = np.meshgrid(xx, yy)\nXfull = np.c_[xx.ravel(), yy.ravel()]\n\nfor index, (name, classifier) in enumerate(classifiers.items()):\n    classifier.fit(X, y)\n\n    y_pred = classifier.predict(X)\n    accuracy = accuracy_score(y, y_pred)\n    print(\"Accuracy (train) for %s: %0.1f%% \" % (name, accuracy * 100))\n\n    # View probabilities:\n    probas = classifier.predict_proba(Xfull)\n    n_classes = np.unique(y_pred).size\n    for k in range(n_classes):\n        plt.subplot(n_classifiers, n_classes, index * n_classes + k + 1)\n        plt.title(\"Class %d\" % k)\n        if k == 0:\n            plt.ylabel(name)\n        imshow_handle = plt.imshow(probas[:, k].reshape((100, 100)),\n                                   extent=(3, 9, 1, 5), origin='lower')\n        plt.xticks(())\n        plt.yticks(())\n        idx = (y_pred == k)\n        if idx.any():\n            plt.scatter(X[idx, 0], X[idx, 1], marker='o', c='w', edgecolor='k')\n\nax = plt.axes([0.15, 0.04, 0.7, 0.05])\nplt.title(\"Probability\")\nplt.colorbar(imshow_handle, cax=ax, orientation='horizontal')\n\nplt.show()"
       ]
     }
   ],

dev/_downloads/auto_examples_python.zip

@@ -3,13 +3,17 @@
 Plot classification probability
 ===============================
 
-Plot the classification probability for different classifiers. We use a 3
-class dataset, and we classify it with a Support Vector classifier, L1
-and L2 penalized logistic regression with either a One-Vs-Rest or multinomial
-setting, and Gaussian process classification.
+Plot the classification probability for different classifiers. We use a 3 class
+dataset, and we classify it with a Support Vector classifier, L1 and L2
+penalized logistic regression with either a One-Vs-Rest or multinomial setting,
+and Gaussian process classification.
 
-The logistic regression is not a multiclass classifier out of the box. As
-a result it can identify only the first class.
+Linear SVC is not a probabilistic classifier by default but it has a built-in
+calibration option enabled in this example (`probability=True`).
+
+The logistic regression with One-Vs-Rest is not a multiclass classifier out of
+the box. As a result it has more trouble in separating class 2 and 3 than the
+other estimators.
 """
 print(__doc__)
 
@@ -19,6 +23,7 @@ class dataset, and we classify it with a Support Vector classifier, L1
 import matplotlib.pyplot as plt
 import numpy as np
 
+from sklearn.metrics import accuracy_score
 from sklearn.linear_model import LogisticRegression
 from sklearn.svm import SVC
 from sklearn.gaussian_process import GaussianProcessClassifier
@@ -31,19 +36,27 @@ class dataset, and we classify it with a Support Vector classifier, L1
 
 n_features = X.shape[1]
 
-C = 1.0
+C = 10
 kernel = 1.0 * RBF([1.0, 1.0])  # for GPC
 
-# Create different classifiers. The logistic regression cannot do
-# multiclass out of the box.
-classifiers = {'L1 logistic': LogisticRegression(C=C, penalty='l1'),
-               'L2 logistic (OvR)': LogisticRegression(C=C, penalty='l2'),
-               'Linear SVC': SVC(kernel='linear', C=C, probability=True,
-                                 random_state=0),
-               'L2 logistic (Multinomial)': LogisticRegression(
-                C=C, solver='lbfgs', multi_class='multinomial'),
-               'GPC': GaussianProcessClassifier(kernel)
-               }
+# Create different classifiers.
+classifiers = {
+    'L1 logistic': LogisticRegression(C=C, penalty='l1',
+                                      solver='saga',
+                                      multi_class='multinomial',
+                                      max_iter=10000),
+    'L2 logistic (Multinomial)': LogisticRegression(C=C, penalty='l2',
+                                                    solver='saga',
+                                                    multi_class='multinomial',
+                                                    max_iter=10000),
+    'L2 logistic (OvR)': LogisticRegression(C=C, penalty='l2',
+                                            solver='saga',
+                                            multi_class='ovr',
+                                            max_iter=10000),
+    'Linear SVC': SVC(kernel='linear', C=C, probability=True,
+                      random_state=0),
+    'GPC': GaussianProcessClassifier(kernel)
+}
 
 n_classifiers = len(classifiers)
 
@@ -59,10 +72,10 @@ class dataset, and we classify it with a Support Vector classifier, L1
     classifier.fit(X, y)
 
     y_pred = classifier.predict(X)
-    classif_rate = np.mean(y_pred.ravel() == y.ravel()) * 100
-    print("classif_rate for %s : %f " % (name, classif_rate))
+    accuracy = accuracy_score(y, y_pred)
+    print("Accuracy (train) for %s: %0.1f%% " % (name, accuracy * 100))
 
-    # View probabilities=
+    # View probabilities:
     probas = classifier.predict_proba(Xfull)
     n_classes = np.unique(y_pred).size
     for k in range(n_classes):

dev/_downloads/plot_classification_probability.ipynb

@@ -26,7 +26,7 @@
       },
       "outputs": [],
       "source": [
-        "# Author: Adam Kleczewski\n# License: BSD 3 clause\n\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.datasets import fetch_openml\nfrom sklearn.multioutput import ClassifierChain\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.multiclass import OneVsRestClassifier\nfrom sklearn.metrics import jaccard_similarity_score\nfrom sklearn.linear_model import LogisticRegression\n\nprint(__doc__)\n\n# Load a multi-label dataset from https://www.openml.org/d/40597\nX, Y = fetch_openml('yeast', version=4, return_X_y=True)\nY = Y == 'TRUE'\nX_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=.2,\n                                                    random_state=0)\n\n# Fit an independent logistic regression model for each class using the\n# OneVsRestClassifier wrapper.\novr = OneVsRestClassifier(LogisticRegression())\novr.fit(X_train, Y_train)\nY_pred_ovr = ovr.predict(X_test)\novr_jaccard_score = jaccard_similarity_score(Y_test, Y_pred_ovr)\n\n# Fit an ensemble of logistic regression classifier chains and take the\n# take the average prediction of all the chains.\nchains = [ClassifierChain(LogisticRegression(), order='random', random_state=i)\n          for i in range(10)]\nfor chain in chains:\n    chain.fit(X_train, Y_train)\n\nY_pred_chains = np.array([chain.predict(X_test) for chain in\n                          chains])\nchain_jaccard_scores = [jaccard_similarity_score(Y_test, Y_pred_chain >= .5)\n                        for Y_pred_chain in Y_pred_chains]\n\nY_pred_ensemble = Y_pred_chains.mean(axis=0)\nensemble_jaccard_score = jaccard_similarity_score(Y_test,\n                                                  Y_pred_ensemble >= .5)\n\nmodel_scores = [ovr_jaccard_score] + chain_jaccard_scores\nmodel_scores.append(ensemble_jaccard_score)\n\nmodel_names = ('Independent',\n               'Chain 1',\n               'Chain 2',\n               'Chain 3',\n               'Chain 4',\n               'Chain 5',\n               'Chain 6',\n               'Chain 7',\n               'Chain 8',\n               'Chain 9',\n               'Chain 10',\n               'Ensemble')\n\nx_pos = np.arange(len(model_names))\n\n# Plot the Jaccard similarity scores for the independent model, each of the\n# chains, and the ensemble (note that the vertical axis on this plot does\n# not begin at 0).\n\nfig, ax = plt.subplots(figsize=(7, 4))\nax.grid(True)\nax.set_title('Classifier Chain Ensemble Performance Comparison')\nax.set_xticks(x_pos)\nax.set_xticklabels(model_names, rotation='vertical')\nax.set_ylabel('Jaccard Similarity Score')\nax.set_ylim([min(model_scores) * .9, max(model_scores) * 1.1])\ncolors = ['r'] + ['b'] * len(chain_jaccard_scores) + ['g']\nax.bar(x_pos, model_scores, alpha=0.5, color=colors)\nplt.tight_layout()\nplt.show()"
+        "# Author: Adam Kleczewski\n# License: BSD 3 clause\n\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.datasets import fetch_openml\nfrom sklearn.multioutput import ClassifierChain\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.multiclass import OneVsRestClassifier\nfrom sklearn.metrics import jaccard_similarity_score\nfrom sklearn.linear_model import LogisticRegression\n\nprint(__doc__)\n\n# Load a multi-label dataset from https://www.openml.org/d/40597\nX, Y = fetch_openml('yeast', version=4, return_X_y=True)\nY = Y == 'TRUE'\nX_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=.2,\n                                                    random_state=0)\n\n# Fit an independent logistic regression model for each class using the\n# OneVsRestClassifier wrapper.\nbase_lr = LogisticRegression(solver='lbfgs')\novr = OneVsRestClassifier(base_lr)\novr.fit(X_train, Y_train)\nY_pred_ovr = ovr.predict(X_test)\novr_jaccard_score = jaccard_similarity_score(Y_test, Y_pred_ovr)\n\n# Fit an ensemble of logistic regression classifier chains and take the\n# take the average prediction of all the chains.\nchains = [ClassifierChain(base_lr, order='random', random_state=i)\n          for i in range(10)]\nfor chain in chains:\n    chain.fit(X_train, Y_train)\n\nY_pred_chains = np.array([chain.predict(X_test) for chain in\n                          chains])\nchain_jaccard_scores = [jaccard_similarity_score(Y_test, Y_pred_chain >= .5)\n                        for Y_pred_chain in Y_pred_chains]\n\nY_pred_ensemble = Y_pred_chains.mean(axis=0)\nensemble_jaccard_score = jaccard_similarity_score(Y_test,\n                                                  Y_pred_ensemble >= .5)\n\nmodel_scores = [ovr_jaccard_score] + chain_jaccard_scores\nmodel_scores.append(ensemble_jaccard_score)\n\nmodel_names = ('Independent',\n               'Chain 1',\n               'Chain 2',\n               'Chain 3',\n               'Chain 4',\n               'Chain 5',\n               'Chain 6',\n               'Chain 7',\n               'Chain 8',\n               'Chain 9',\n               'Chain 10',\n               'Ensemble')\n\nx_pos = np.arange(len(model_names))\n\n# Plot the Jaccard similarity scores for the independent model, each of the\n# chains, and the ensemble (note that the vertical axis on this plot does\n# not begin at 0).\n\nfig, ax = plt.subplots(figsize=(7, 4))\nax.grid(True)\nax.set_title('Classifier Chain Ensemble Performance Comparison')\nax.set_xticks(x_pos)\nax.set_xticklabels(model_names, rotation='vertical')\nax.set_ylabel('Jaccard Similarity Score')\nax.set_ylim([min(model_scores) * .9, max(model_scores) * 1.1])\ncolors = ['r'] + ['b'] * len(chain_jaccard_scores) + ['g']\nax.bar(x_pos, model_scores, alpha=0.5, color=colors)\nplt.tight_layout()\nplt.show()"
       ]
     }
   ],

dev/_downloads/plot_classification_probability.py

@@ -54,14 +54,15 @@
 
 # Fit an independent logistic regression model for each class using the
 # OneVsRestClassifier wrapper.
-ovr = OneVsRestClassifier(LogisticRegression())
+base_lr = LogisticRegression(solver='lbfgs')
+ovr = OneVsRestClassifier(base_lr)
 ovr.fit(X_train, Y_train)
 Y_pred_ovr = ovr.predict(X_test)
 ovr_jaccard_score = jaccard_similarity_score(Y_test, Y_pred_ovr)
 
 # Fit an ensemble of logistic regression classifier chains and take the
 # take the average prediction of all the chains.
-chains = [ClassifierChain(LogisticRegression(), order='random', random_state=i)
+chains = [ClassifierChain(base_lr, order='random', random_state=i)
           for i in range(10)]
 for chain in chains:
     chain.fit(X_train, Y_train)