update to regressor

VIDA-NYU · Sep 1, 2024 · 7942cd0 · 7942cd0
1 parent 7b57db4
commit 7942cd0
Show file tree

Hide file tree

Showing 3 changed files with 786 additions and 751 deletions.
diff --git a/alpha_automl/hyperparameter_tuning/smac.py b/alpha_automl/hyperparameter_tuning/smac.py
@@ -41,11 +41,14 @@ def gen_pipeline(config, pipeline):
 
         if step_type == 'COLUMN_TRANSFORMER':
             transformers = []
-            for trans_name, _, trans_index in step_obj.__dict__['transformers']:
+            for trans_name, trans_obj_ori, trans_index in step_obj.__dict__['transformers']:
                 trans_prim_name = trans_name.split('-')[0]
-                trans_obj = create_object(trans_prim_name, get_primitive_params(config, trans_prim_name))
-                transformers.append((trans_name, trans_obj, trans_index))
-                step_obj.__dict__['transformers'] = transformers
+                if "alpha_automl.builtin_primitives.math_features" in trans_prim_name:
+                    transformers.append((trans_name, trans_obj_ori, trans_index))
+                else:
+                    trans_obj = create_object(trans_prim_name, get_primitive_params(config, trans_prim_name))
+                    transformers.append((trans_name, trans_obj, trans_index))
+            step_obj.__dict__['transformers'] = transformers
             new_pipeline.steps.append([step_name, create_object(step_name, step_obj.__dict__)])
         else:
             new_pipeline.steps.append([step_name, create_object(step_name, get_primitive_params(config, step_name))])

diff --git a/alpha_automl/hyperparameter_tuning/smac_parameters.json b/alpha_automl/hyperparameter_tuning/smac_parameters.json
@@ -274,8 +274,229 @@
     },
     "sklearn.preprocessing.PolynomialFeatures": {},
     "alpha_automl.wrapper_primitives.llm_feature_engine.LLMFeatureGenerator": {},
-    "feature_engine.creation.math_features.MathFeatures-sum": {},
-    "feature_engine.creation.math_features.MathFeatures-mean": {},
-    "feature_engine.creation.math_features.MathFeatures-prod": {},
-    "feature_engine.creation.math_features.MathFeatures-std": {}
+    "alpha_automl.builtin_primitives.math_features.MathFeaturesProd": {},
+    "alpha_automl.builtin_primitives.math_features.MathFeaturesMean": {},
+    "alpha_automl.builtin_primitives.math_features.MathFeaturesSum": {},
+    "alpha_automl.builtin_primitives.math_features.MathFeaturesStd": {},
+    "sklearn.ensemble.ExtraTreesRegressor": {
+        "n_estimators": {
+            "type": "Integer",
+            "value": [
+                200,
+                2000
+            ],
+            "default": 400
+        },
+        "max_features": {
+            "type": "Float",
+            "value": [
+                0.1,
+                1.0
+            ],
+            "default": 1.0
+        },
+        "max_leaf_nodes": {
+            "type": "Integer",
+            "value": [
+                80,
+                10000
+            ],
+            "default": 80
+        },
+        "criterion": {
+            "type": "Categorical",
+            "value": [
+                "entropy",
+                "gini"
+            ],
+            "default": "entropy"
+        }
+    },
+    "sklearn.ensemble.RandomForestRegressor": {
+        "n_estimators": {
+            "type": "Integer",
+            "value": [
+                200,
+                2000
+            ],
+            "default": 400
+        },
+        "max_features": {
+            "type": "Float",
+            "value": [
+                0.1,
+                1.0
+            ],
+            "default": 1.0
+        },
+        "criterion": {
+            "type": "Categorical",
+            "value": [
+                "entropy",
+                "gini"
+            ],
+            "default": "entropy"
+        }
+    },
+    "xgboost.XGBRegressor": {
+        "n_estimators": {
+            "type": "Integer",
+            "value": [
+                200,
+                2000
+            ],
+            "default": 400
+        },
+        "max_leaves": {
+            "type": "Integer",
+            "value": [
+                80,
+                10000
+            ],
+            "default": 80
+        },
+        "min_child_weight": {
+            "type": "Float",
+            "value": [
+                0.001,
+                0.1
+            ],
+            "default": 0.1
+        },
+        "learning_rate": {
+            "type": "Float",
+            "value": [
+                0.01,
+                0.3
+            ],
+            "default": 0.1
+        },
+        "subsample": {
+            "type": "Float",
+            "value": [
+                0.5,
+                1.0
+            ],
+            "default": 1.0
+        },
+        "colsample_bylevel": {
+            "type": "Float",
+            "value": [
+                0.5,
+                1.0
+            ],
+            "default": 1.0
+        },
+        "colsample_bytree": {
+            "type": "Float",
+            "value": [
+                0.5,
+                1.0
+            ],
+            "default": 1.0
+        }
+    },
+    "lightgbm.LGBMRegressor": {
+        "n_estimators": {
+            "type": "Integer",
+            "value": [
+                200,
+                2000
+            ],
+            "default": 400
+        },
+        "num_leaves": {
+            "type": "Integer",
+            "value": [
+                80,
+                10000
+            ],
+            "default": 80
+        },
+        "min_child_samples": {
+            "type": "Integer",
+            "value": [
+                20,
+                100
+            ],
+            "default": 20
+        },
+        "learning_rate": {
+            "type": "Float",
+            "value": [
+                0.001,
+                0.3
+            ],
+            "default": 0.1
+        },
+        "log_max_bin": {
+            "type": "Integer",
+            "value": [
+                6,
+                10
+            ],
+            "default": 8
+        },
+        "colsample_bytree": {
+            "type": "Float",
+            "value": [
+                0.3,
+                1.0
+            ],
+            "default": 1.0
+        },
+        "verbose": {
+            "type": "Constant",
+            "value": -1,
+            "default": -1
+        }
+    },
+    "catboost.CatBoostRegressor": {
+        "logging_level": {
+            "type": "Constant",
+            "value": "Silent",
+            "default": "Silent"
+        },
+        "learning_rate": {
+            "type": "Float",
+            "value": [
+                0.01,
+                0.3
+            ],
+            "default": 0.1
+        },
+        "depth": {
+            "type": "Integer",
+            "value": [
+                1,
+                6
+            ],
+            "default": 6
+        },
+        "l2_leaf_reg": {
+            "type": "Float",
+            "value": [
+                1,
+                10
+            ],
+            "default": 3
+        },
+        "border_count": {
+            "type": "Integer",
+            "value": [
+                32,
+                255
+            ],
+            "default": 254
+        },
+        "auto_class_weights": {
+            "type": "Categorical",
+            "value": [
+                "None",
+                "Balanced",
+                "SqrtBalanced"
+            ],
+            "default": "None"
+        }
+    }
 }