diff --git a/python/treelite/sklearn/importer.py b/python/treelite/sklearn/importer.py
index f198078a..02270375 100644
--- a/python/treelite/sklearn/importer.py
+++ b/python/treelite/sklearn/importer.py
@@ -67,9 +67,17 @@ def import_model(sklearn_model):
     ----
     For :py:class:`~sklearn.ensemble.IsolationForest`, the loaded model will calculate the outlier
     score using the standardized ratio as proposed in the original reference,
-    which matches with :py:meth:`~sklearn.ensemble.IsolationForest._compute_chunked_score_samples`
+    which matches with :py:meth:`~sklearn.ensemble.IsolationForest.score_samples`
     but is a bit different from :py:meth:`~sklearn.ensemble.IsolationForest.decision_function`.
 
+    More precisely, the following relation holds:
+
+    .. code-block:: python
+
+        treelite.gtil.predict(tl_model, X) == -clf.score_samples(X)
+        # clf is an IsolationForest
+        # tl_model is a Treelite representation of clf
+
     Parameters
     ----------
     sklearn_model : object of type \
diff --git a/tests/python/test_sklearn_integration.py b/tests/python/test_sklearn_integration.py
index 67ff13ea..4b6f8436 100644
--- a/tests/python/test_sklearn_integration.py
+++ b/tests/python/test_sklearn_integration.py
@@ -187,7 +187,7 @@ def test_skl_converter_iforest(dataset):
         random_state=0,
     )
     clf.fit(X)
-    expected_pred = clf._compute_chunked_score_samples(X)  # pylint: disable=W0212
+    expected_pred = -clf.score_samples(X)
     expected_pred = expected_pred.reshape((-1, 1, 1))
 
     tl_model = treelite.sklearn.import_model(clf)