Stacking all the models

stacking.py

from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import MultinomialNB
from xgboost import XGBClassifier
from sklearn.svm import SVC

# Define all models 
lr = SGDClassifier(loss = 'log', alpha = 0.1, penalty = 'elasticnet')
svm = SVC(C = 10, kernel = 'poly', degree = 2, probability = True)
nb = MultinomialNB(alpha = 10000, class_prior = [0.5, 0.5])
knn = KNeighborsClassifier(n_neighbors = 7, weights = 'distance', n_jobs = -1)
rf = RandomForestClassifier(n_estimators = 250, min_samples_split = 5, max_depth = 15,  n_jobs = -1)
xgb = XGBClassifier(n_estimators = 100, learning_rate = 0.3, max_depth = 1, n_jobs = -1)

model_dict = dict(zip(['LR', 'SVM', 'NB', 'KNN', 'RF', 'XGB'], [lr, svm, nb, knn, rf, xgb]))

for model_name, model in model_dict.items():
    print('Training {}'.format(model_name))
    model.fit(train_features, y_train)

model_weights = {   'LR' : 0.9,
                    'SVM' : 0.9,
                    'NB' : 0.8,
                    'KNN' : 0.75,
                    'RF' : 0.75,
                    'XGB' : 0.6,
                    'simple_nn' : 0.7
}

y_pred_prob = 0

for model_name, model in model_dict.items():
    y_pred_prob += (model.predict_proba(test_features)[:,1] * model_weights[model_name])

y_pred_prob += (simple_nn.predict(test_features.todense()).ravel() * model_weights['simple_nn'])
y_pred_prob /= sum(model_weights.values())

print_model_metrics(y_test, y_pred_prob)