Merge pull request scikit-learn#4961 from mrphilroth/issue4959

[MRG] Allowing for different cluster stds

Author: ogrisel

sklearn/datasets/samples_generator.py

@@ -745,6 +745,9 @@ def make_blobs(n_samples=100, n_features=2, centers=3, cluster_std=1.0,
         centers = check_array(centers)
         n_features = centers.shape[1]
 
+    if isinstance(cluster_std, numbers.Real):
+        cluster_std = np.ones(len(centers)) * cluster_std
+
     X = []
     y = []
 
@@ -754,8 +757,8 @@ def make_blobs(n_samples=100, n_features=2, centers=3, cluster_std=1.0,
     for i in range(n_samples % n_centers):
         n_samples_per_center[i] += 1
 
-    for i, n in enumerate(n_samples_per_center):
-        X.append(centers[i] + generator.normal(scale=cluster_std,
+    for i, (n, std) in enumerate(zip(n_samples_per_center, cluster_std)):
+        X.append(centers[i] + generator.normal(scale=std,
                                                size=(n, n_features)))
         y += [i] * n
 

sklearn/datasets/tests/test_samples_generator.py

@@ -208,13 +208,16 @@ def test_make_regression_multitarget():
 
 
 def test_make_blobs():
-    X, y = make_blobs(n_samples=50, n_features=2,
-                      centers=[[0.0, 0.0], [1.0, 1.0], [0.0, 1.0]],
-                      random_state=0)
+    cluster_stds = np.array([0.05, 0.2, 0.4])
+    cluster_centers = np.array([[0.0, 0.0], [1.0, 1.0], [0.0, 1.0]])
+    X, y = make_blobs(random_state=0, n_samples=50, n_features=2,
+                      centers=cluster_centers, cluster_std=cluster_stds)
 
     assert_equal(X.shape, (50, 2), "X shape mismatch")
     assert_equal(y.shape, (50,), "y shape mismatch")
     assert_equal(np.unique(y).shape, (3,), "Unexpected number of blobs")
+    for i, (ctr, std) in enumerate(zip(cluster_centers, cluster_stds)):
+        assert_almost_equal((X[y == i] - ctr).std(), std, 1, "Unexpected std")
 
 
 def test_make_friedman1():