Merge pull request #104 from VIDA-NYU/reduce_search_space

Author: roquelopez

alpha_automl/pipeline_synthesis/pipeline_builder.py

@@ -34,7 +34,7 @@ def change_default_hyperparams(primitive_object):
     elif isinstance(primitive_object, SimpleImputer):
         primitive_object.set_params(strategy='most_frequent', keep_empty_features=True)
     elif isinstance(primitive_object, AdaBoostClassifier):
-        primitive_object.set_params(algorithm='SAMME')
+        primitive_object.set_params(algorithm='SAMME.R')
 
 
 def extract_estimators(pipeline_primitives, all_primitives):

alpha_automl/resource/primitives_hierarchy.json

@@ -52,48 +52,15 @@
         "sklearn.ensemble.VotingRegressor"
     ],
     "CLASSIFIER": [
-        "sklearn.discriminant_analysis.LinearDiscriminantAnalysis",
-        "sklearn.discriminant_analysis.QuadraticDiscriminantAnalysis",
         "sklearn.ensemble.ExtraTreesClassifier",
-        "sklearn.ensemble.GradientBoostingClassifier",
         "sklearn.ensemble.RandomForestClassifier",
-        "sklearn.naive_bayes.BernoulliNB",
-        "sklearn.naive_bayes.GaussianNB",
-        "sklearn.naive_bayes.MultinomialNB",
-        "sklearn.neighbors.KNeighborsClassifier",
-        "sklearn.linear_model.LogisticRegression",
-        "sklearn.linear_model.PassiveAggressiveClassifier",
-        "sklearn.linear_model.SGDClassifier",
-        "sklearn.svm.LinearSVC",
-        "sklearn.svm.SVC",
-        "sklearn.tree.DecisionTreeClassifier",
         "xgboost.XGBClassifier",
         "lightgbm.LGBMClassifier",
         "catboost.CatBoostClassifier"
     ],
     "REGRESSOR": [
-        "sklearn.linear_model.ARDRegression",
-        "sklearn.tree.DecisionTreeRegressor",
         "sklearn.ensemble.ExtraTreesRegressor",
-        "sklearn.gaussian_process.GaussianProcessRegressor",
-        "sklearn.ensemble.GradientBoostingRegressor",
-        "sklearn.neighbors.KNeighborsRegressor",
-        "sklearn.linear_model.Lars",
-        "sklearn.linear_model.Lasso",
-        "sklearn.linear_model.LassoCV",
-        "sklearn.svm.LinearSVR",
-        "sklearn.linear_model.PassiveAggressiveRegressor",
         "sklearn.ensemble.RandomForestRegressor",
-        "sklearn.linear_model.Ridge",
-        "sklearn.linear_model.SGDRegressor",
-        "sklearn.svm.SVR",
-        "sklearn.linear_model.BayesianRidge",
-        "sklearn.linear_model.ElasticNet",
-        "sklearn.linear_model.HuberRegressor",
-        "sklearn.linear_model.LinearRegression",
-        "sklearn.linear_model.RANSACRegressor",
-        "sklearn.linear_model.RidgeCV",
-        "sklearn.linear_model.TheilSenRegressor",
         "xgboost.XGBRegressor",
         "lightgbm.LGBMRegressor",
         "catboost.CatBoostRegressor"

alpha_automl/scorer.py

@@ -7,7 +7,7 @@
 from alpha_automl.utils import RANDOM_SEED, hide_logs
 from sklearn.metrics import accuracy_score, f1_score, jaccard_score, precision_score, recall_score,\
     max_error, mean_absolute_error, mean_squared_error, mean_squared_log_error, median_absolute_error, r2_score,\
-    adjusted_mutual_info_score, rand_score, mutual_info_score, normalized_mutual_info_score
+    adjusted_mutual_info_score, rand_score, mutual_info_score, normalized_mutual_info_score, roc_auc_score
 from alpha_automl.pipeline import Pipeline
 from alpha_automl.primitive_loader import PRIMITIVE_TYPES
 
@@ -19,6 +19,7 @@
     'f1_score': f1_score,
     'precision_score': precision_score,
     'recall_score': recall_score,
+    'roc_auc_score': roc_auc_score,
     'jaccard_score': jaccard_score,
     # Regression metrics
     'max_error': max_error,
@@ -41,6 +42,7 @@
     f1_score: 'ascending',
     precision_score: 'ascending',
     recall_score: 'ascending',
+    roc_auc_score: 'ascending', 
     jaccard_score: 'ascending',
     # Regression metrics
     max_error: 'descending',