Add feature generator category

VIDA-NYU · May 7, 2024 · bf419fd · bf419fd
1 parent 98ddbaa
commit bf419fd
Show file tree

Hide file tree

Showing 2 changed files with 58 additions and 54 deletions.
diff --git a/alpha_automl/resource/base_grammar.bnf b/alpha_automl/resource/base_grammar.bnf
@@ -1,15 +1,16 @@
-S -> CLASSIFICATION_TASK | REGRESSION_TASK | CLUSTERING_TASK | TIME_SERIES_FORECAST_TASK | SEMISUPERVISED_TASK
-CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER CLASSIFICATION_ENSEMBLER
-REGRESSION_TASK -> IMPUTER ENCODERS FEATURE_SELECTOR FEATURE_SCALER REGRESSOR REGRESSION_ENSEMBLER
-CLUSTERING_TASK -> IMPUTER ENCODERS FEATURE_SCALER FEATURE_SELECTOR CLUSTERER
+S -> CLASSIFICATION_TASK | REGRESSION_TASK | CLUSTERING_TASK | TIME_SERIES_FORECAST_TASK | SEMISUPERVISED_TASK | NA_TASK
+CLASSIFICATION_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR CLASSIFIER CLASSIFICATION_ENSEMBLER
+REGRESSION_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR REGRESSOR REGRESSION_ENSEMBLER
+CLUSTERING_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER FEATURE_SELECTOR CLUSTERER
 TIME_SERIES_FORECAST_TASK -> IMPUTER TIME_SERIES_FORECASTER | REGRESSION_TASK
-SEMISUPERVISED_TASK -> IMPUTER ENCODERS FEATURE_SCALER SEMISUPERVISED_CLASSIFIER
+SEMISUPERVISED_TASK -> IMPUTER ENCODERS FEATURE_GENERATOR FEATURE_SCALER SEMISUPERVISED_CLASSIFIER
 NA_TASK -> CLASSIFICATION_TASK | REGRESSION_TASK | SEMISUPERVISED_TASK
 ENCODERS -> TEXT_ENCODER DATETIME_ENCODER CATEGORICAL_ENCODER IMAGE_ENCODER
 CLASSIFICATION_ENSEMBLER -> CLASSIFICATION_SINGLE_ENSEMBLER | CLASSIFIER CLASSIFIER CLASSIFICATION_MULTI_ENSEMBLER | E
 REGRESSION_ENSEMBLER -> REGRESSION_SINGLE_ENSEMBLER | REGRESSOR REGRESSOR REGRESSION_MULTI_ENSEMBLER | E
 SEMISUPERVISED_CLASSIFIER -> CLASSIFIER SEMISUPERVISED_SELFTRAINER | SEMISUPERVISED_LABELPROPAGATOR
 IMPUTER -> 'primitive_terminal'
+FEATURE_GENERATOR -> 'primitive_terminal' | 'E'
 FEATURE_SCALER -> 'primitive_terminal' | 'E'
 FEATURE_SELECTOR -> 'primitive_terminal' | 'E'
 TEXT_ENCODER -> 'primitive_terminal'

diff --git a/alpha_automl/resource/primitives_hierarchy.json b/alpha_automl/resource/primitives_hierarchy.json
@@ -1,7 +1,56 @@
 {
+    "IMPUTER": [
+        "sklearn.impute.SimpleImputer"
+    ],
     "CATEGORICAL_ENCODER": [
         "sklearn.preprocessing.OneHotEncoder"
     ],
+    "DATETIME_ENCODER": [
+        "sklearn.preprocessing.OrdinalEncoder",
+	    "alpha_automl.builtin_primitives.datetime_encoder.CyclicalFeature",
+        "alpha_automl.builtin_primitives.datetime_encoder.Datetime64ExpandEncoder",
+        "alpha_automl.builtin_primitives.datetime_encoder.DummyEncoder"
+    ],
+    "TEXT_ENCODER": [
+        "sklearn.feature_extraction.text.CountVectorizer",
+        "sklearn.feature_extraction.text.TfidfVectorizer"
+    ],
+    "IMAGE_ENCODER": [
+        "alpha_automl.builtin_primitives.image_encoder.RGB2GrayTransformer",
+        "alpha_automl.builtin_primitives.image_encoder.HogTransformer"
+    ],
+    "FEATURE_GENERATOR": [
+        "sklearn.preprocessing.PolynomialFeatures"
+    ],
+    "FEATURE_SCALER": [
+        "sklearn.preprocessing.MaxAbsScaler",
+        "sklearn.preprocessing.RobustScaler",
+        "sklearn.preprocessing.StandardScaler"
+    ],
+    "FEATURE_SELECTOR": [
+        "sklearn.feature_selection.GenericUnivariateSelect",
+        "sklearn.feature_selection.SelectPercentile",
+        "sklearn.feature_selection.SelectKBest"
+    ],
+    "COLUMN_TRANSFORMER": [
+        "sklearn.compose.ColumnTransformer"
+    ],
+    "CLASSIFICATION_SINGLE_ENSEMBLER": [
+        "sklearn.ensemble.AdaBoostClassifier",
+        "sklearn.ensemble.BaggingClassifier"
+    ],
+    "CLASSIFICATION_MULTI_ENSEMBLER": [
+        "sklearn.ensemble.StackingClassifier",
+        "sklearn.ensemble.VotingClassifier"
+    ],
+    "REGRESSION_SINGLE_ENSEMBLER": [
+        "sklearn.ensemble.AdaBoostRegressor",
+        "sklearn.ensemble.BaggingRegressor"
+    ],
+    "REGRESSION_MULTI_ENSEMBLER": [
+        "sklearn.ensemble.StackingRegressor",
+        "sklearn.ensemble.VotingRegressor"
+    ],
     "CLASSIFIER": [
         "sklearn.discriminant_analysis.LinearDiscriminantAnalysis",
         "sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis",
@@ -22,29 +71,6 @@
         "lightgbm.LGBMClassifier",
         "catboost.CatBoostClassifier"
     ],
-    "CLUSTERER": [
-        "sklearn.cluster.KMeans",
-        "sklearn.cluster.AgglomerativeClustering"
-    ],
-    "DATETIME_ENCODER": [
-        "sklearn.preprocessing.OrdinalEncoder",
-	    "alpha_automl.builtin_primitives.datetime_encoder.CyclicalFeature",
-        "alpha_automl.builtin_primitives.datetime_encoder.Datetime64ExpandEncoder",
-        "alpha_automl.builtin_primitives.datetime_encoder.DummyEncoder"
-    ],
-    "FEATURE_SCALER": [
-        "sklearn.preprocessing.MaxAbsScaler",
-        "sklearn.preprocessing.RobustScaler",
-        "sklearn.preprocessing.StandardScaler"
-    ],
-    "FEATURE_SELECTOR": [
-        "sklearn.feature_selection.GenericUnivariateSelect",
-        "sklearn.feature_selection.SelectPercentile",
-        "sklearn.feature_selection.SelectKBest"
-    ],
-    "IMPUTER": [
-        "sklearn.impute.SimpleImputer"
-    ],
     "REGRESSOR": [
         "sklearn.linear_model.ARDRegression",
         "sklearn.tree.DecisionTreeRegressor",
@@ -72,16 +98,9 @@
         "lightgbm.LGBMRegressor",
         "catboost.CatBoostRegressor"
     ],
-    "TEXT_ENCODER": [
-        "sklearn.feature_extraction.text.CountVectorizer",
-        "sklearn.feature_extraction.text.TfidfVectorizer"
-    ],
-    "IMAGE_ENCODER": [
-        "alpha_automl.builtin_primitives.image_encoder.RGB2GrayTransformer",
-        "alpha_automl.builtin_primitives.image_encoder.HogTransformer"
-    ],
-    "COLUMN_TRANSFORMER": [
-        "sklearn.compose.ColumnTransformer"
+    "CLUSTERER": [
+        "sklearn.cluster.KMeans",
+        "sklearn.cluster.AgglomerativeClustering"
     ],
     "TIME_SERIES_FORECASTER": [
         "alpha_automl.builtin_primitives.time_series_forecasting.ArimaEstimator",
@@ -96,21 +115,5 @@
     "SEMISUPERVISED_LABELPROPAGATOR": [
         "alpha_automl.builtin_primitives.semisupervised_classifier.SkLabelSpreading",
         "alpha_automl.builtin_primitives.semisupervised_classifier.SkLabelPropagation"
-    ],
-    "CLASSIFICATION_SINGLE_ENSEMBLER": [
-        "sklearn.ensemble.AdaBoostClassifier",
-        "sklearn.ensemble.BaggingClassifier"
-    ],
-    "CLASSIFICATION_MULTI_ENSEMBLER": [
-        "sklearn.ensemble.StackingClassifier",
-        "sklearn.ensemble.VotingClassifier"
-    ],
-    "REGRESSION_SINGLE_ENSEMBLER": [
-        "sklearn.ensemble.AdaBoostRegressor",
-        "sklearn.ensemble.BaggingRegressor"
-    ],
-    "REGRESSION_MULTI_ENSEMBLER": [
-        "sklearn.ensemble.StackingRegressor",
-        "sklearn.ensemble.VotingRegressor"
     ]
 }