Implement flexible switching between classification and regression

create.py

+"""
+Functions that create a machine learning model from training data
+"""
+
 import os
 import sys
 import logging

model_creator.py

@@ -103,15 +103,7 @@ def get_cv_error(clf,feats,scores):
 
     return results
 
-def extract_features_and_generate_model_predictors(predictor_set, type=util_functions.AlgorithmTypes.regression):
-    if(algorithm not in [util_functions.AlgorithmTypes.regression, util_functions.AlgorithmTypes.classification]):
-        algorithm = util_functions.AlgorithmTypes.regression
-
-    f = predictor_extractor.PredictorExtractor()
-    f.initialize_dictionaries(predictor_set)
-
-    train_feats = f.gen_feats(predictor_set)
-
+def get_algorithms(type):
     if type == util_functions.AlgorithmTypes.classification:
         clf = sklearn.ensemble.GradientBoostingClassifier(n_estimators=100, learn_rate=.05,
             max_depth=4, random_state=1,min_samples_leaf=3)
@@ -122,7 +114,19 @@ def extract_features_and_generate_model_predictors(predictor_set, type=util_func
             max_depth=4, random_state=1,min_samples_leaf=3)
         clf2=sklearn.ensemble.GradientBoostingRegressor(n_estimators=100, learn_rate=.05,
             max_depth=4, random_state=1,min_samples_leaf=3)
+    return clf, clf2
+
 
+def extract_features_and_generate_model_predictors(predictor_set, type=util_functions.AlgorithmTypes.regression):
+    if(algorithm not in [util_functions.AlgorithmTypes.regression, util_functions.AlgorithmTypes.classification]):
+        algorithm = util_functions.AlgorithmTypes.regression
+
+    f = predictor_extractor.PredictorExtractor()
+    f.initialize_dictionaries(predictor_set)
+
+    train_feats = f.gen_feats(predictor_set)
+
+    clf,clf2 = get_algorithms(type)
     cv_error_results=get_cv_error(clf2,train_feats,predictor_set._target)
 
     try:
@@ -153,16 +157,17 @@ def extract_features_and_generate_model(essays,additional_array=None):
         if(additional_array.shape[0]==train_feats.shape[0]):
             train_feats=numpy.concatenate((train_feats,additional_array),axis=1)
 
-    clf = sklearn.ensemble.GradientBoostingClassifier(n_estimators=100, learn_rate=.05,
-        max_depth=4, random_state=1,min_samples_leaf=3)
+    set_score = numpy.asarray(essays._score, dtype=numpy.int)
+    if len(util_functions.f7(list(set_score)))>5:
+        type = util_functions.AlgorithmTypes.regression
+    else:
+        type = util_functions.AlgorithmTypes.classification
 
-    clf2=sklearn.ensemble.GradientBoostingClassifier(n_estimators=100, learn_rate=.05,
-        max_depth=4, random_state=1,min_samples_leaf=3)
+    clf,clf2 = get_algorithms(type)
 
     cv_error_results=get_cv_error(clf2,train_feats,essays._score)
 
     try:
-        set_score = numpy.asarray(essays._score, dtype=numpy.int)
         clf.fit(train_feats, set_score)
     except ValueError:
         log.exception("Not enough classes (0,1,etc) in sample.")