FIX mutual_info_regression when X is of integer dtype (scikit-learn#26748)

Author: Charlie-XIAO

doc/whats_new/v1.4.rst

@@ -93,3 +93,10 @@ TODO: update at the time of the release.
   when using a custom initialization. The default value of this parameter will change
   from `None` to `auto` in version 1.6.
   :pr:`26634` by :user:`Alexandre Landeau <AlexL>` and :user:`Alexandre Vigny <avigny>`.
+
+
+:mod:`sklearn.feature_selection`
+................................
+
+- |Fix| :func:`feature_selection.mutual_info_regression` now correctly computes the
+  result when `X` is of integer dtype. :pr:`26748` by :user:`Yao Xiao <Charlie-XIAO>`.

sklearn/feature_selection/_mutual_info.py

@@ -280,15 +280,12 @@ def _estimate_mi(
 
     rng = check_random_state(random_state)
     if np.any(continuous_mask):
-        if copy:
-            X = X.copy()
-
+        X = X.astype(np.float64, copy=copy)
         X[:, continuous_mask] = scale(
             X[:, continuous_mask], with_mean=False, copy=False
         )
 
         # Add small noise to continuous features as advised in Kraskov et. al.
-        X = X.astype(np.float64, copy=False)
         means = np.maximum(1, np.mean(np.abs(X[:, continuous_mask]), axis=0))
         X[:, continuous_mask] += (
             1e-10

sklearn/feature_selection/tests/test_mutual_info.py

@@ -236,3 +236,18 @@ def test_mutual_information_symmetry_classif_regression(correlated, global_rando
     )
 
     assert mi_classif == pytest.approx(mi_regression)
+
+
+def test_mutual_info_regression_X_int_dtype(global_random_seed):
+    """Check that results agree when X is integer dtype and float dtype.
+
+    Non-regression test for Issue #26696.
+    """
+    rng = np.random.RandomState(global_random_seed)
+    X = rng.randint(100, size=(100, 10))
+    X_float = X.astype(np.float64, copy=True)
+    y = rng.randint(100, size=100)
+
+    expected = mutual_info_regression(X_float, y, random_state=global_random_seed)
+    result = mutual_info_regression(X, y, random_state=global_random_seed)
+    assert_allclose(result, expected)