DOC roc_auc_score and average_precision_score explicit about binary input (#9557)

jrbourbeau · jnothman · commit 6f42105ac1a8 · 2017-08-22T10:02:03.000+10:00
diff --git a/sklearn/metrics/ranking.py b/sklearn/metrics/ranking.py
@@ -116,7 +116,7 @@ def average_precision_score(y_true, y_score, average="macro",
     Parameters
     ----------
     y_true : array, shape = [n_samples] or [n_samples, n_classes]
-        True binary labels in binary label indicators.
+        True binary labels (either {0, 1} or {-1, 1}).
 
     y_score : array, shape = [n_samples] or [n_samples, n_classes]
         Target scores, can either be probability estimates of the positive
@@ -200,7 +200,7 @@ def roc_auc_score(y_true, y_score, average="macro", sample_weight=None):
     Parameters
     ----------
     y_true : array, shape = [n_samples] or [n_samples, n_classes]
-        True binary labels in binary label indicators.
+        True binary labels (either {0, 1} or {-1, 1}).
 
     y_score : array, shape = [n_samples] or [n_samples, n_classes]
         Target scores, can either be probability estimates of the positive
@@ -438,8 +438,8 @@ def roc_curve(y_true, y_score, pos_label=None, sample_weight=None,
     ----------
 
     y_true : array, shape = [n_samples]
-        True binary labels in range {0, 1} or {-1, 1}.  If labels are not
-        binary, pos_label should be explicitly given.
+        True binary labels. If labels are not either {-1, 1} or {0, 1}, then
+        pos_label should be explicitly given.
 
     y_score : array, shape = [n_samples]
         Target scores, can either be probability estimates of the positive
