glemaitre
diff --git a/Collapse file
‎sklearn/utils/estimator_checks.py‎
Copy file name to clipboardExpand all lines: sklearn/utils/estimator_checks.py
+6-3Lines changed: 6 additions & 3 deletions b/Collapse file
‎sklearn/utils/estimator_checks.py‎
Copy file name to clipboardExpand all lines: sklearn/utils/estimator_checks.py
+6-3Lines changed: 6 additions & 3 deletions
@@ -2907,16 +2907,19 @@ def check_decision_proba_consistency(name, estimator_orig):
     centers = [(2, 2), (4, 4)]
     X, y = make_blobs(n_samples=100, random_state=0, n_features=4,
                       centers=centers, cluster_std=1.0, shuffle=True)
-    X_test = np.random.randn(20, 2) + 4
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2,
+                                                        random_state=0)
     estimator = clone(estimator_orig)
 
     if (hasattr(estimator, "decision_function") and
             hasattr(estimator, "predict_proba")):
 
-        estimator.fit(X, y)
+        estimator.fit(X_train, y_train)
         # Since the link function from decision_function() to predict_proba()
         # is sometimes not precise enough (typically expit), we round to the
-        # 10th decimal to avoid numerical issues.
+        # 10th decimal to avoid numerical issues: we compare the rank
+        # with deterministic ties rather than get platform specific rank
+        # inversions in case of machine level differences.
         a = estimator.predict_proba(X_test)[:, 1].round(decimals=10)
         b = estimator.decision_function(X_test).round(decimals=10)
         assert_array_equal(rankdata(a), rankdata(b))