Update

feature_importance/01_run_importance_local_simulations.py → feature_importance/01_run_importance_local_ablation_regression.py

@@ -117,28 +117,34 @@ def compare_estimators(estimators: List[ModelConfig],
                     'splitting_strategy': splitting_strategy
                 }
                 start = time.time()
-                local_fi_score = fi_est.cls(X_test, y_test, copy.deepcopy(est), **fi_est.kwargs)
+                if fi_est.name in ["MDI_local_sub_stumps_evaluate", "MDI_local_all_stumps_evaluate", "LFI_absolute_sum_evaluate",
+                                   "MDI_local_sub_stumps_evaluate_without_raw", "MDI_local_all_stumps_evaluate_without_raw",
+                                   "LFI_absolute_sum_evaluate_without_raw"]:
+                    local_fi_score = fi_est.cls(X_train, y_train, X_test, y_test, copy.deepcopy(est), **fi_est.kwargs)
+                else:
+                    local_fi_score = fi_est.cls(X_test, y_test, copy.deepcopy(est), **fi_est.kwargs)
                 end = time.time()
                 metric_results['fi_time'] = end - start
                 feature_importance_list.append(local_fi_score)
                 support_df = pd.DataFrame({"var": np.arange(len(support)),
                                            "true_support": support,
                                            "cor_with_signal": x_cor})
                 metric_results['fi_scores'] = support_df
+
                 start = time.time()
-                if np.max(support) != np.min(support):
-                    y_pred = est.predict(X_test)
-                    metric_results['MSE_before_ablation'] = mean_squared_error(y_test, y_pred)
-                    imp_vals = copy.deepcopy(local_fi_score)
-                    imp_vals[imp_vals == float("-inf")] = -sys.maxsize - 1
-                    imp_vals[imp_vals == float("inf")] = sys.maxsize - 1
-                    for i in range(X_test.shape[1]):
-                        if fi_est.ascending:
-                            ablation_X_test = ablation(X_test, imp_vals, "max", i+1, seed)
-                        else:
-                            ablation_X_test = ablation(X_test, imp_vals, "min", i+1, seed)
-                        metric_results[f'MSE_after_ablation_{i+1}'] = mean_squared_error(y_test, est.predict(ablation_X_test))
+                y_pred = est.predict(X_test)
+                metric_results['MSE_before_ablation'] = mean_squared_error(y_test, y_pred)
+                imp_vals = copy.deepcopy(local_fi_score)
+                imp_vals[imp_vals == float("-inf")] = -sys.maxsize - 1
+                imp_vals[imp_vals == float("inf")] = sys.maxsize - 1
+                for i in range(X_test.shape[1]):
+                    if fi_est.ascending:
+                        ablation_X_test = ablation(X_test, imp_vals, "max", i+1, seed)
+                    else:
+                        ablation_X_test = ablation(X_test, imp_vals, "min", i+1, seed)
+                    metric_results[f'MSE_after_ablation_{i+1}'] = mean_squared_error(y_test, est.predict(ablation_X_test))
                 end = time.time()
+
                 metric_results['ablation_time'] = end - start
                 metric_results['test_size'] = X_test.shape[0]
                 print(f"data_size: {X_test.shape[0]}, fi: {fi_est.name}, done with time: {end - start}")