Merge pull request #7 from mrahim/predict_proba_estimators

ENH: add predict proba estimator

Author: mrahim

stlearn/stacking.py

@@ -115,6 +115,9 @@ class StackingClassifier(BaseEstimator, ClassifierMixin, TransformerMixin):
     features_indices : list of indexers
         Index epxressions to be applied on the columns of X_stacked.
         Can be slices, lists of intgers or bool.
+
+    n_jobs : int (default: 1)
+        The number of jobs to run in parallel (across estimators).
     """
 
     def __init__(self, estimators,
@@ -260,6 +263,27 @@ def predict_estimators(self, X):
             for x, clf in zip(X_list, self.estimators))
         return np.array(predictions_).T
 
+    def predict_proba_estimators(self, X):
+        """Predict class label probability for samples in X for each estimators.
+
+        Parameters
+        ----------
+        X : {array-like, sparse matrix}, shape = (n_samples, n_features)
+            The multi-input samples.
+
+        Returns
+        -------
+        C : array, shape = (n_samples, n_classes, n_estimators)
+            Predicted class label per sample and estimators.
+        """
+        _check_Xy(self, X)
+        X_list = _split_features(X, self.feature_indices)
+        predictions_ = Parallel(n_jobs=self.n_jobs)(
+            delayed(_predict_proba_estimator)(clf, x)
+            for x, clf in zip(X_list, self.estimators))
+        return np.transpose(np.array(predictions_),
+                            (1, 2, 0))
+
     def score_estimators(self, X, y):
         """Returns the mean accuracy for each estimators.
 

stlearn/tests/test_stacking.py

@@ -70,6 +70,11 @@ def test_stacking_essentials():
     proba = stacking.predict_proba(X_stacked)
     assert_array_equal(proba.sum(1), np.ones_like(proba[:, 1]))
 
+    proba_estimators = stacking.predict_proba_estimators(X_stacked)
+    for proba in np.transpose(proba_estimators, (2, 0, 1)):
+        assert_array_equal(proba.sum(1),
+                           np.ones_like(proba[:, 1]))
+
     score = stacking.score(X_stacked, y)
     assert_true(np.isscalar(score))