kf = StratifiedKFold(n_splits=5, shuffle=True, random_state=RND_SEED)
top_score = 10
top_classifier = None

#from sklearn.grid_search import GridSearchCV
for train_index, test_index in kf.split(X_train, y_train):
    for n_estimators in [1, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100]:
        for max_depth in [1, 3, 10]:
            for loss in ["deviance", "exponential"]:
                gbc = GradientBoostingClassifier(n_estimators=n_estimators, max_depth=max_depth,loss=loss)
                X_train2, X_test2 = X_train.as_matrix()[train_index], X_train.as_matrix()[test_index]
                y_train2, y_test2 = y_train.as_matrix()[train_index], y_train.as_matrix()[test_index]
                gbc.fit(X_train2, y_train2)
                current_score = gbc.score(X_test2, y_test2)
                if current_score < top_score:
                    top_score = current_score
                    top_classifier = gbc

from sklearn.ensemble import AdaBoostClassifier

kf = StratifiedKFold(n_splits=5, shuffle=True, random_state=RND_SEED)
top_score = 10
top_classifier = None

#from sklearn.grid_search import GridSearchCV
for train_index, test_index in kf.split(X_train, y_train):
    for n_estimators in [1, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100]:
                adb = AdaBoostClassifier(n_estimators=n_estimators)
                X_train2, X_test2 = X_train.as_matrix()[train_index], X_train.as_matrix()[test_index]
                y_train2, y_test2 = y_train.as_matrix()[train_index], y_train.as_matrix()[test_index]
                adb.fit(X_train2, y_train2)
                current_score = adb.score(X_test2, y_test2)
                if current_score < top_score:
                    top_score = current_score
                    top_classifier = adb

#Ada