Merge pull request scikit-learn#2895 from kaushik94/master

 supervised.py with updated documentation

Author: agramfort

sklearn/metrics/cluster/supervised.py

@@ -596,8 +596,9 @@ def adjusted_mutual_info_score(labels_true, labels_pred):
 
     Returns
     -------
-    ami: float
-       score between 0.0 and 1.0. 1.0 stands for perfectly complete labeling
+    ami: float(upperlimited by 1.0)
+       The AMI takes a value of 1 when the two partitions 
+       are identical and a value of 0 is expected for random partitions.
 
     See also
     --------