FIX AffinityPropagation assigning multiple clusters for equal points (#28121)

Charlie-XIAO · web-flow · commit 783643525060 · 2024-01-18T11:44:26.000Z
diff --git a/doc/whats_new/v1.4.rst b/doc/whats_new/v1.4.rst
@@ -2,6 +2,25 @@
 
 .. currentmodule:: sklearn
 
+.. _changes_1_4_1:
+
+Version 1.4.1
+=============
+
+**In Development**
+
+Changelog
+---------
+
+:mod:`sklearn.cluster`
+......................
+
+- |Fix| :class:`cluster.AffinityPropagation` now avoids assigning multiple different
+  clusters for equal points.
+  :pr:`28121` by :user:`Pietro Peterlongo <pietroppeter>` and
+  :user:`Yao Xiao <Charlie-XIAO>`.
+
+
 .. _changes_1_4:
 
 Version 1.4.0
diff --git a/sklearn/cluster/_affinity_propagation.py b/sklearn/cluster/_affinity_propagation.py
@@ -53,7 +53,7 @@ def _affinity_propagation(
             "All samples have mutually equal similarities. "
             "Returning arbitrary cluster center(s)."
         )
-        if preference.flat[0] >= S.flat[n_samples - 1]:
+        if preference.flat[0] > S.flat[n_samples - 1]:
             return (
                 (np.arange(n_samples), np.arange(n_samples), 0)
                 if return_n_iter
diff --git a/sklearn/cluster/tests/test_affinity_propagation.py b/sklearn/cluster/tests/test_affinity_propagation.py
@@ -308,3 +308,14 @@ def test_sparse_input_for_fit_predict(csr_container):
     X = csr_container(rng.randint(0, 2, size=(5, 5)))
     labels = af.fit_predict(X)
     assert_array_equal(labels, (0, 1, 1, 2, 3))
+
+
+def test_affinity_propagation_equal_points():
+    """Make sure we do not assign multiple clusters to equal points.
+
+    Non-regression test for:
+    https://github.com/scikit-learn/scikit-learn/pull/20043
+    """
+    X = np.zeros((8, 1))
+    af = AffinityPropagation(affinity="euclidean", damping=0.5, random_state=42).fit(X)
+    assert np.all(af.labels_ == 0)

Original file line number	Diff line number	Diff line change
`@@ -53,7 +53,7 @@ def _affinity_propagation(`
`53`	`53`	`"All samples have mutually equal similarities. "`
`54`	`54`	`"Returning arbitrary cluster center(s)."`
`55`	`55`	`)`
`56`		`- if preference.flat[0] >= S.flat[n_samples - 1]:`
	`56`	`+ if preference.flat[0] > S.flat[n_samples - 1]:`
`57`	`57`	`return (`
`58`	`58`	`(np.arange(n_samples), np.arange(n_samples), 0)`
`59`	`59`	`if return_n_iter`