rename ari_score adjusted_rand_score

doc/modules/classes.rst

@@ -295,7 +295,7 @@ Clustering metrics
    :toctree: generated/
    :template: function.rst
 
-   metrics.ari_score
+   metrics.adjusted_rand_score
    metrics.homogeneity_completeness_v_measure
    metrics.homogeneity_score
    metrics.completeness_score

doc/modules/clustering.rst

@@ -358,34 +358,34 @@ chance normalization**::
   >>> labels_true = [0, 0, 0, 1, 1, 1]
   >>> labels_pred = [0, 0, 1, 1, 2, 2]
 
-  >>> metrics.ari_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
+  >>> metrics.adjusted_rand_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
   0.24...
 
 One can permute 0 and 1 in the predicted labels and rename `2` by `3` and get
 the same score::
 
   >>> labels_pred = [1, 1, 0, 0, 3, 3]
-  >>> metrics.ari_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
+  >>> metrics.adjusted_rand_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
   0.24...
 
-Furthermore, :func:`ari_score` is **symmetric**: swapping the argument
+Furthermore, :func:`adjusted_rand_score` is **symmetric**: swapping the argument
 does not change the score. It can thus be used as a **consensus
 measure**::
 
-  >>> metrics.ari_score(labels_pred, labels_true)  # doctest: +ELLIPSIS
+  >>> metrics.adjusted_rand_score(labels_pred, labels_true)  # doctest: +ELLIPSIS
   0.24...
 
 Perfect labeling is scored 1.0::
 
   >>> labels_pred = labels_true[:]
-  >>> metrics.ari_score(labels_true, labels_pred)
+  >>> metrics.adjusted_rand_score(labels_true, labels_pred)
   1.0
 
 Bad (e.g. independent labelings) have negative or close to 0.0 scores::
 
   >>> labels_true = [0, 1, 2, 0, 3, 4, 5, 1]
   >>> labels_pred = [1, 1, 0, 0, 2, 2, 2, 2]
-  >>> metrics.ari_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
+  >>> metrics.adjusted_rand_score(labels_true, labels_pred)  # doctest: +ELLIPSIS
   -0.12...
 
 

examples/cluster/plot_adjusted_for_chance_measures.py

@@ -56,7 +56,7 @@ def uniform_labelings_scores(score_func, n_samples, n_clusters_range,
     return scores
 
 score_funcs = [
-    metrics.ari_score,
+    metrics.adjusted_rand_score,
     metrics.v_measure_score,
 ]
 

examples/cluster/plot_affinity_propagation.py

@@ -38,7 +38,8 @@ print 'Estimated number of clusters: %d' % n_clusters_
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels)
 print "Completeness: %0.3f" % metrics.completeness_score(labels_true, labels)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels)
-print "Adjusted Rand Index: %0.3f" % metrics.ari_score(labels_true, labels)
+print "Adjusted Rand Index: %0.3f" % \
+    metrics.adjusted_rand_score(labels_true, labels)
 
 ##############################################################################
 # Plot result

examples/cluster/plot_dbscan.py

@@ -39,7 +39,8 @@ print 'Estimated number of clusters: %d' % n_clusters_
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels_true, labels)
 print "Completeness: %0.3f" % metrics.completeness_score(labels_true, labels)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels_true, labels)
-print "Adjusted Rand Index: %0.3f" % metrics.ari_score(labels_true, labels)
+print "Adjusted Rand Index: %0.3f" % \
+    metrics.adjusted_rand_score(labels_true, labels)
 
 ##############################################################################
 # Plot result

examples/cluster/plot_kmeans_digits.py

@@ -44,7 +44,8 @@ print "Inertia: %f" % km.inertia_
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels, km.labels_)
 print "Completeness: %0.3f" % metrics.completeness_score(labels, km.labels_)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels, km.labels_)
-print "Adjusted Rand Index: %0.3f" % metrics.ari_score(labels, km.labels_)
+print "Adjusted Rand Index: %0.3f" % \
+    metrics.adjusted_rand_score(labels, km.labels_)
 print
 
 print "Raw k-means with random centroid init..."
@@ -55,7 +56,8 @@ print "Inertia: %f" % km.inertia_
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels, km.labels_)
 print "Completeness: %0.3f" % metrics.completeness_score(labels, km.labels_)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels, km.labels_)
-print "Adjusted Rand Index: %0.3f" % metrics.ari_score(labels, km.labels_)
+print "Adjusted Rand Index: %0.3f" % \
+    metrics.adjusted_rand_score(labels, km.labels_)
 print
 
 print "Raw k-means with PCA-based centroid init..."
@@ -69,7 +71,8 @@ print "Inertia: %f" % km.inertia_
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels, km.labels_)
 print "Completeness: %0.3f" % metrics.completeness_score(labels, km.labels_)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels, km.labels_)
-print "Adjusted Rand Index: %0.3f" % metrics.ari_score(labels, km.labels_)
+print "Adjusted Rand Index: %0.3f" % \
+    metrics.adjusted_rand_score(labels, km.labels_)
 print
 
 # Plot k-means++ form on a 2D plot using PCA

examples/document_clustering.py

@@ -85,5 +85,6 @@ print
 print "Homogeneity: %0.3f" % metrics.homogeneity_score(labels, mbkm.labels_)
 print "Completeness: %0.3f" % metrics.completeness_score(labels, mbkm.labels_)
 print "V-measure: %0.3f" % metrics.v_measure_score(labels, mbkm.labels_)
-print "Adjusted Rand-Index: %.3f" % metrics.ari_score(labels, mbkm.labels_)
+print "Adjusted Rand-Index: %.3f" % \
+    metrics.adjusted_rand_score(labels, mbkm.labels_)
 print

sklearn/metrics/__init__.py

@@ -9,7 +9,7 @@ from .metrics import confusion_matrix, roc_curve, auc, precision_score, \
                 precision_recall_curve, explained_variance_score, r2_score, \
                 zero_one, mean_square_error, hinge_loss
 
-from .cluster import ari_score
+from .cluster import adjusted_rand_score
 from .cluster import homogeneity_completeness_v_measure
 from .cluster import homogeneity_score
 from .cluster import completeness_score

sklearn/metrics/cluster.py

@@ -12,6 +12,7 @@ from scipy import comb
 
 import numpy as np
 
+
 # the exact version if faster for k == 2: use it by default globally in
 # this module instead of the float approximate variant
 def comb2(n):
@@ -39,7 +40,7 @@ def check_clusterings(labels_true, labels_pred):
 
 # clustering measures
 
-def ari_score(labels_true, labels_pred):
+def adjusted_rand_score(labels_true, labels_pred):
     """Rand index adjusted for chance
 
     The Rand Index computes a similarity measure between two clusterings
@@ -59,7 +60,7 @@ def ari_score(labels_true, labels_pred):
 
     ARI is a symmetric measure::
 
-        ari_score(a, b) == ari_score(b, a)
+        adjusted_rand_score(a, b) == adjusted_rand_score(b, a)
 
     Parameters
     ----------
@@ -80,28 +81,28 @@ def ari_score(labels_true, labels_pred):
 
     Perfectly maching labelings have a score of 1 even
 
-      >>> from sklearn.metrics.cluster import ari_score
-      >>> ari_score([0, 0, 1, 1], [0, 0, 1, 1])
+      >>> from sklearn.metrics.cluster import adjusted_rand_score
+      >>> adjusted_rand_score([0, 0, 1, 1], [0, 0, 1, 1])
       1.0
-      >>> ari_score([0, 0, 1, 1], [1, 1, 0, 0])
+      >>> adjusted_rand_score([0, 0, 1, 1], [1, 1, 0, 0])
       1.0
 
     Labelings that assign all classes members to the same clusters
     are complete be not always pure, hence penalized::
 
-      >>> ari_score([0, 0, 1, 2], [0, 0, 1, 1])     # doctest: +ELLIPSIS
+      >>> adjusted_rand_score([0, 0, 1, 2], [0, 0, 1, 1])  # doctest: +ELLIPSIS
       0.57...
 
     ARI is symmetric, so labelings that have pure clusters with members
     coming from the same classes but unnecessary splits are penalized::
 
-      >>> ari_score([0, 0, 1, 1], [0, 0, 1, 2])     # doctest: +ELLIPSIS
+      >>> adjusted_rand_score([0, 0, 1, 1], [0, 0, 1, 2])  # doctest: +ELLIPSIS
       0.57...
 
     If classes members are completely split across different clusters, the
     assignment is totally incomplete, hence the ARI is very low::
 
-      >>> ari_score([0, 0, 0, 0], [0, 1, 2, 3])
+      >>> adjusted_rand_score([0, 0, 0, 0], [0, 1, 2, 3])
       0.0
 
     References

sklearn/metrics/tests/test_cluster.py

 import numpy as np
 
-from sklearn.metrics.cluster import ari_score
+from sklearn.metrics.cluster import adjusted_rand_score
 from sklearn.metrics.cluster import homogeneity_score
 from sklearn.metrics.cluster import completeness_score
 from sklearn.metrics.cluster import v_measure_score
@@ -12,12 +12,13 @@ from numpy.testing import assert_array_almost_equal
 
 
 score_funcs = [
-    ari_score,
+    adjusted_rand_score,
     homogeneity_score,
     completeness_score,
     v_measure_score,
 ]
 
+
 def assert_raise_message(exception, message, callable, *args, **kwargs):
     """Helper function to test error messages in exceptions"""
     try:
@@ -97,8 +98,8 @@ def test_non_consicutive_labels():
     assert_almost_equal(c, 0.42, 2)
     assert_almost_equal(v, 0.52, 2)
 
-    ari_1 = ari_score([0, 0, 0, 1, 1, 1], [0, 1, 0, 1, 2, 2])
-    ari_2 = ari_score([0, 0, 0, 1, 1, 1], [0, 4, 0, 4, 2, 2])
+    ari_1 = adjusted_rand_score([0, 0, 0, 1, 1, 1], [0, 1, 0, 1, 2, 2])
+    ari_2 = adjusted_rand_score([0, 0, 0, 1, 1, 1], [0, 4, 0, 4, 2, 2])
     assert_almost_equal(ari_1, 0.24, 2)
     assert_almost_equal(ari_2, 0.24, 2)
 
@@ -123,7 +124,7 @@ def test_adjustment_for_chance():
     n_runs = 10
 
     scores = uniform_labelings_scores(
-        ari_score, n_samples, n_clusters_range, n_runs)
+        adjusted_rand_score, n_samples, n_clusters_range, n_runs)
 
     max_abs_scores = np.abs(scores).max(axis=1)
     assert_array_almost_equal(max_abs_scores, [0.02, 0.03, 0.03, 0.02], 2)