Pushing the docs to dev/ for branch: master, commit 726fa36f2556e0d604d85a1de48ba56a8b6550db

Author: sklearn-ci

dev/_downloads/auto_examples_jupyter.zip

@@ -15,7 +15,7 @@
       "cell_type": "markdown",
       "metadata": {},
       "source": [
-        "\n# Imputing missing values before building an estimator\n\n\nMissing values can be replaced by the mean, the median or the most frequent\nvalue using the basic ``SimpleImputer``.\nThe median is a more robust estimator for data with high magnitude variables\nwhich could dominate results (otherwise known as a 'long tail').\n\nAnother option is the ``ChainedImputer``. This uses round-robin linear\nregression, treating every variable as an output in turn. The version\nimplemented assumes Gaussian (output) variables. If your features are obviously\nnon-Normal, consider transforming them to look more Normal so as to improve\nperformance.\n\n"
+        "\n# Imputing missing values before building an estimator\n\n\nMissing values can be replaced by the mean, the median or the most frequent\nvalue using the basic :func:`sklearn.impute.SimpleImputer`.\nThe median is a more robust estimator for data with high magnitude variables\nwhich could dominate results (otherwise known as a 'long tail').\n\nAnother option is the :func:`sklearn.impute.ChainedImputer`. This uses\nround-robin linear regression, treating every variable as an output in\nturn. The version implemented assumes Gaussian (output) variables. If your\nfeatures are obviously non-Normal, consider transforming them to look more\nNormal so as to improve performance.\n\nIn addition of using an imputing method, we can also keep an indication of the\nmissing information using :func:`sklearn.impute.MissingIndicator` which might\ncarry some information.\n\n"
       ]
     },
     {
@@ -26,7 +26,7 @@
       },
       "outputs": [],
       "source": [
-        "import numpy as np\nimport matplotlib.pyplot as plt\n\nfrom sklearn.datasets import load_diabetes\nfrom sklearn.datasets import load_boston\nfrom sklearn.ensemble import RandomForestRegressor\nfrom sklearn.pipeline import Pipeline\nfrom sklearn.impute import SimpleImputer, ChainedImputer\nfrom sklearn.model_selection import cross_val_score\n\nrng = np.random.RandomState(0)\n\n\ndef get_results(dataset):\n    X_full, y_full = dataset.data, dataset.target\n    n_samples = X_full.shape[0]\n    n_features = X_full.shape[1]\n\n    # Estimate the score on the entire dataset, with no missing values\n    estimator = RandomForestRegressor(random_state=0, n_estimators=100)\n    full_scores = cross_val_score(estimator, X_full, y_full,\n                                  scoring='neg_mean_squared_error')\n\n    # Add missing values in 75% of the lines\n    missing_rate = 0.75\n    n_missing_samples = int(np.floor(n_samples * missing_rate))\n    missing_samples = np.hstack((np.zeros(n_samples - n_missing_samples,\n                                          dtype=np.bool),\n                                 np.ones(n_missing_samples,\n                                         dtype=np.bool)))\n    rng.shuffle(missing_samples)\n    missing_features = rng.randint(0, n_features, n_missing_samples)\n\n    # Estimate the score after replacing missing values by 0\n    X_missing = X_full.copy()\n    X_missing[np.where(missing_samples)[0], missing_features] = 0\n    y_missing = y_full.copy()\n    estimator = RandomForestRegressor(random_state=0, n_estimators=100)\n    zero_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                         scoring='neg_mean_squared_error')\n\n    # Estimate the score after imputation (mean strategy) of the missing values\n    X_missing = X_full.copy()\n    X_missing[np.where(missing_samples)[0], missing_features] = 0\n    y_missing = y_full.copy()\n    estimator = Pipeline([(\"imputer\", SimpleImputer(missing_values=0,\n                                                    strategy=\"mean\")),\n                          (\"forest\", RandomForestRegressor(random_state=0,\n                                                           n_estimators=100))])\n    mean_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                         scoring='neg_mean_squared_error')\n\n    # Estimate the score after chained imputation of the missing values\n    estimator = Pipeline([(\"imputer\", ChainedImputer(missing_values=0,\n                                                     random_state=0)),\n                          (\"forest\", RandomForestRegressor(random_state=0,\n                                                           n_estimators=100))])\n    chained_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                            scoring='neg_mean_squared_error')\n\n    return ((full_scores.mean(), full_scores.std()),\n            (zero_impute_scores.mean(), zero_impute_scores.std()),\n            (mean_impute_scores.mean(), mean_impute_scores.std()),\n            (chained_impute_scores.mean(), chained_impute_scores.std()))\n\n\nresults_diabetes = np.array(get_results(load_diabetes()))\nmses_diabetes = results_diabetes[:, 0] * -1\nstds_diabetes = results_diabetes[:, 1]\n\nresults_boston = np.array(get_results(load_boston()))\nmses_boston = results_boston[:, 0] * -1\nstds_boston = results_boston[:, 1]\n\nn_bars = len(mses_diabetes)\nxval = np.arange(n_bars)\n\nx_labels = ['Full data',\n            'Zero imputation',\n            'Mean Imputation',\n            'Chained Imputation']\ncolors = ['r', 'g', 'b', 'orange']\n\n# plot diabetes results\nplt.figure(figsize=(12, 6))\nax1 = plt.subplot(121)\nfor j in xval:\n    ax1.barh(j, mses_diabetes[j], xerr=stds_diabetes[j],\n             color=colors[j], alpha=0.6, align='center')\n\nax1.set_title('Imputation Techniques with Diabetes Data')\nax1.set_xlim(left=np.min(mses_diabetes) * 0.9,\n             right=np.max(mses_diabetes) * 1.1)\nax1.set_yticks(xval)\nax1.set_xlabel('MSE')\nax1.invert_yaxis()\nax1.set_yticklabels(x_labels)\n\n# plot boston results\nax2 = plt.subplot(122)\nfor j in xval:\n    ax2.barh(j, mses_boston[j], xerr=stds_boston[j],\n             color=colors[j], alpha=0.6, align='center')\n\nax2.set_title('Imputation Techniques with Boston Data')\nax2.set_yticks(xval)\nax2.set_xlabel('MSE')\nax2.invert_yaxis()\nax2.set_yticklabels([''] * n_bars)\n\nplt.show()"
+        "import numpy as np\nimport matplotlib.pyplot as plt\n\nfrom sklearn.datasets import load_diabetes\nfrom sklearn.datasets import load_boston\nfrom sklearn.ensemble import RandomForestRegressor\nfrom sklearn.pipeline import make_pipeline, make_union\nfrom sklearn.impute import SimpleImputer, ChainedImputer, MissingIndicator\nfrom sklearn.model_selection import cross_val_score\n\nrng = np.random.RandomState(0)\n\n\ndef get_results(dataset):\n    X_full, y_full = dataset.data, dataset.target\n    n_samples = X_full.shape[0]\n    n_features = X_full.shape[1]\n\n    # Estimate the score on the entire dataset, with no missing values\n    estimator = RandomForestRegressor(random_state=0, n_estimators=100)\n    full_scores = cross_val_score(estimator, X_full, y_full,\n                                  scoring='neg_mean_squared_error')\n\n    # Add missing values in 75% of the lines\n    missing_rate = 0.75\n    n_missing_samples = int(np.floor(n_samples * missing_rate))\n    missing_samples = np.hstack((np.zeros(n_samples - n_missing_samples,\n                                          dtype=np.bool),\n                                 np.ones(n_missing_samples,\n                                         dtype=np.bool)))\n    rng.shuffle(missing_samples)\n    missing_features = rng.randint(0, n_features, n_missing_samples)\n\n    # Estimate the score after replacing missing values by 0\n    X_missing = X_full.copy()\n    X_missing[np.where(missing_samples)[0], missing_features] = 0\n    y_missing = y_full.copy()\n    estimator = RandomForestRegressor(random_state=0, n_estimators=100)\n    zero_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                         scoring='neg_mean_squared_error')\n\n    # Estimate the score after imputation (mean strategy) of the missing values\n    X_missing = X_full.copy()\n    X_missing[np.where(missing_samples)[0], missing_features] = 0\n    y_missing = y_full.copy()\n    estimator = make_pipeline(\n        make_union(SimpleImputer(missing_values=0, strategy=\"mean\"),\n                   MissingIndicator(missing_values=0)),\n        RandomForestRegressor(random_state=0, n_estimators=100))\n    mean_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                         scoring='neg_mean_squared_error')\n\n    # Estimate the score after chained imputation of the missing values\n    estimator = make_pipeline(\n        make_union(ChainedImputer(missing_values=0, random_state=0),\n                   MissingIndicator(missing_values=0)),\n        RandomForestRegressor(random_state=0, n_estimators=100))\n    chained_impute_scores = cross_val_score(estimator, X_missing, y_missing,\n                                            scoring='neg_mean_squared_error')\n\n    return ((full_scores.mean(), full_scores.std()),\n            (zero_impute_scores.mean(), zero_impute_scores.std()),\n            (mean_impute_scores.mean(), mean_impute_scores.std()),\n            (chained_impute_scores.mean(), chained_impute_scores.std()))\n\n\nresults_diabetes = np.array(get_results(load_diabetes()))\nmses_diabetes = results_diabetes[:, 0] * -1\nstds_diabetes = results_diabetes[:, 1]\n\nresults_boston = np.array(get_results(load_boston()))\nmses_boston = results_boston[:, 0] * -1\nstds_boston = results_boston[:, 1]\n\nn_bars = len(mses_diabetes)\nxval = np.arange(n_bars)\n\nx_labels = ['Full data',\n            'Zero imputation',\n            'Mean Imputation',\n            'Chained Imputation']\ncolors = ['r', 'g', 'b', 'orange']\n\n# plot diabetes results\nplt.figure(figsize=(12, 6))\nax1 = plt.subplot(121)\nfor j in xval:\n    ax1.barh(j, mses_diabetes[j], xerr=stds_diabetes[j],\n             color=colors[j], alpha=0.6, align='center')\n\nax1.set_title('Imputation Techniques with Diabetes Data')\nax1.set_xlim(left=np.min(mses_diabetes) * 0.9,\n             right=np.max(mses_diabetes) * 1.1)\nax1.set_yticks(xval)\nax1.set_xlabel('MSE')\nax1.invert_yaxis()\nax1.set_yticklabels(x_labels)\n\n# plot boston results\nax2 = plt.subplot(122)\nfor j in xval:\n    ax2.barh(j, mses_boston[j], xerr=stds_boston[j],\n             color=colors[j], alpha=0.6, align='center')\n\nax2.set_title('Imputation Techniques with Boston Data')\nax2.set_yticks(xval)\nax2.set_xlabel('MSE')\nax2.invert_yaxis()\nax2.set_yticklabels([''] * n_bars)\n\nplt.show()"
       ]
     }
   ],

dev/_downloads/auto_examples_python.zip

@@ -4,15 +4,19 @@
 ====================================================
 
 Missing values can be replaced by the mean, the median or the most frequent
-value using the basic ``SimpleImputer``.
+value using the basic :func:`sklearn.impute.SimpleImputer`.
 The median is a more robust estimator for data with high magnitude variables
 which could dominate results (otherwise known as a 'long tail').
 
-Another option is the ``ChainedImputer``. This uses round-robin linear
-regression, treating every variable as an output in turn. The version
-implemented assumes Gaussian (output) variables. If your features are obviously
-non-Normal, consider transforming them to look more Normal so as to improve
-performance.
+Another option is the :func:`sklearn.impute.ChainedImputer`. This uses
+round-robin linear regression, treating every variable as an output in
+turn. The version implemented assumes Gaussian (output) variables. If your
+features are obviously non-Normal, consider transforming them to look more
+Normal so as to improve performance.
+
+In addition of using an imputing method, we can also keep an indication of the
+missing information using :func:`sklearn.impute.MissingIndicator` which might
+carry some information.
 """
 
 import numpy as np
@@ -21,8 +25,8 @@
 from sklearn.datasets import load_diabetes
 from sklearn.datasets import load_boston
 from sklearn.ensemble import RandomForestRegressor
-from sklearn.pipeline import Pipeline
-from sklearn.impute import SimpleImputer, ChainedImputer
+from sklearn.pipeline import make_pipeline, make_union
+from sklearn.impute import SimpleImputer, ChainedImputer, MissingIndicator
 from sklearn.model_selection import cross_val_score
 
 rng = np.random.RandomState(0)
@@ -60,18 +64,18 @@ def get_results(dataset):
     X_missing = X_full.copy()
     X_missing[np.where(missing_samples)[0], missing_features] = 0
     y_missing = y_full.copy()
-    estimator = Pipeline([("imputer", SimpleImputer(missing_values=0,
-                                                    strategy="mean")),
-                          ("forest", RandomForestRegressor(random_state=0,
-                                                           n_estimators=100))])
+    estimator = make_pipeline(
+        make_union(SimpleImputer(missing_values=0, strategy="mean"),
+                   MissingIndicator(missing_values=0)),
+        RandomForestRegressor(random_state=0, n_estimators=100))
     mean_impute_scores = cross_val_score(estimator, X_missing, y_missing,
                                          scoring='neg_mean_squared_error')
 
     # Estimate the score after chained imputation of the missing values
-    estimator = Pipeline([("imputer", ChainedImputer(missing_values=0,
-                                                     random_state=0)),
-                          ("forest", RandomForestRegressor(random_state=0,
-                                                           n_estimators=100))])
+    estimator = make_pipeline(
+        make_union(ChainedImputer(missing_values=0, random_state=0),
+                   MissingIndicator(missing_values=0)),
+        RandomForestRegressor(random_state=0, n_estimators=100))
     chained_impute_scores = cross_val_score(estimator, X_missing, y_missing,
                                             scoring='neg_mean_squared_error')