Fix TreeExplainer instanciation when a tree in LightGBM Classifier mo…

…del has n_features_in_tree = 1 (#310) * Fix bug at instanciation when a tree has only one feature * Modifty comments * Correction using n_interpolation_size instead of _n_features_in_tree * Add minimal test * Fix styling * ran pre-commit --------- Co-authored-by: mmschlk <[email protected]>
mmschlk · Feb 5, 2025 · 4b04a2f · 4b04a2f
1 parent 0e27619
commit 4b04a2f
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 2 deletions.
diff --git a/shapiq/explainer/tree/treeshapiq.py b/shapiq/explainer/tree/treeshapiq.py
@@ -407,7 +407,13 @@ def _init_summary_polynomials(self):
                 interaction_order=order, n_features=self._n_features_in_tree
             )
             self.subset_ancestors_store[order] = subset_ancestors
-            self.D_store[order] = np.polynomial.chebyshev.chebpts2(self.n_interpolation_size)
+
+            # If the tree has only one feature, we assign a default value of 0
+            if self.n_interpolation_size == 1:
+                self.D_store[order] = np.array([0])
+            else:
+                self.D_store[order] = np.polynomial.chebyshev.chebpts2(self.n_interpolation_size)
+
             self.D_powers_store[order] = self._cache(self.D_store[order])
             if self._index in ("SV", "SII", "k-SII"):
                 self.Ns_store[order] = self._get_N(self.D_store[order])

diff --git a/tests/tests_explainer/tests_tree_explainer/test_tree_treeshapiq.py b/tests/tests_explainer/tests_tree_explainer/test_tree_treeshapiq.py
@@ -3,7 +3,7 @@
 import numpy as np
 import pytest
 
-from shapiq.explainer.tree import TreeModel, TreeSHAPIQ
+from shapiq.explainer.tree import TreeExplainer, TreeModel, TreeSHAPIQ
 
 
 def test_init(dt_clf_model, background_clf_data):
@@ -132,3 +132,29 @@ def test_edge_case_params():
     # test with max_order = 0
     with pytest.raises(ValueError):
         _ = TreeSHAPIQ(model=tree_model, max_order=0)
+
+
+def test_no_bug_with_one_feature_tree():
+    # create the dataset
+    X = np.array(
+        [
+            [1, 1, 1, 1],
+            [1, 1, 1, 2],
+            [2, 1, 1, 1],
+            [3, 2, 1, 1],
+        ]
+    )
+
+    # Define simple one feature tree
+    tree = {
+        "children_left": np.array([1, -1, -1]),
+        "children_right": np.array([2, -1, -1]),
+        "features": np.array([0, -2, -2]),
+        "thresholds": np.array([2.5, -2, -2]),
+        "values": np.array([0.5, 0.0, 1]),
+        "node_sample_weight": np.array([14, 5, 9]),
+    }
+    tree = TreeModel(**tree)
+    explainer = TreeExplainer(model=tree, index="SV", max_order=1)
+    shapley_values = explainer.explain(X[2])
+    print(shapley_values)