def main():
train_path = "../data/churn_train.csv"
test_path = "../data/churn_test.csv"
dc_train = DataCleaning(train_path)
dc_test = DataCleaning(test_path)
X_train, y_train = dc_train.clean()
X_test, y_test = dc_test.clean()
dc_train_reg = DataCleaning(train_path)
dc_test_reg = DataCleaning(test_path)
X_train_reg, y_train_reg = dc_train_reg.clean(regression=True)
X_test_reg, y_test_reg = dc_test_reg.clean(regression=True)
train_col_names = dc_train.get_column_names()
train_col_names_reg = dc_train_reg.get_column_names()
rf = RandomForestClassifier
gb = GradientBoostingClassifier
logr = LogisticRegression
pipe = Pipeline([rf, gb])
pipe.fit_predict(X_train, y_train)
pipe.print_cv_results(train_col_names, X_train, y_train)
pipe2 = Pipeline([logr])
pipe2.fit_predict(X_train_reg, y_train_reg)
pipe2.print_cv_results(train_col_names_reg, X_train_reg, y_train_reg)
plot_rocs([pipe, pipe2], [[X_train, y_train], [X_train_reg, y_train_reg]])
test_scores = pipe.score(X_test, y_test)
def main():
train_path = "data/data.json"
#test_path = "data/test.csv"
dc_train = DataCleaning(train_path)
#dc_test = DataCleaning(test_path)
X_train, y_train = dc_train.clean()
#X_test, y_test = dc_test.clean()
# dc_train_reg = DataCleaning(train_path)
# dc_test_reg = DataCleaning(test_path)
# X_train_reg, y_train_reg = dc_train_reg.clean(regression=True)
# X_test_reg, y_test_reg = dc_test_reg.clean(regression=True)
train_col_names = dc_train.get_column_names()
# train_col_names_reg = dc_train_reg.get_column_names()
rf = RandomForestClassifier
gb = GradientBoostingClassifier
logr = LogisticRegression
svm_model = svm.SVC
pipe = Pipeline([gb])
pipe.fit_predict(X_train, y_train)
pipe.print_cv_results(train_col_names, X_train, y_train)
with open('model.pkl', 'w') as f:
pickle.dump(pipe.trained_models[1], f)