df = pd.read_csv('gym_churn.csv')

df_train, df_valid = train_test_split(df, test_size=0.25, random_state=12345)

features_train = df_train.drop(['Churn',
                                'gender',
                                'Phone',
                                #'Contract_period',
                                #'Age',
                                #'Avg_additional_charges_total',
                                #'Month_to_end_contract',
                                #'Lifetime',
                                #'Avg_class_frequency_total',
                                #'Avg_class_frequency_current_month'
                               ], axis = 1)
target_train = df_train['Churn']

features_valid = df_valid.drop(['Churn',
                                'gender',
                                'Phone',
                                #'Contract_period',
                                #'Age',
                                #'Avg_additional_charges_total',
                                #'Month_to_end_contract',
                                #'Lifetime',
                                #'Avg_class_frequency_total',
                                #'Avg_class_frequency_current_month'
                               ], axis = 1)
target_valid = df_valid['Churn']

for estim in range(5, 70, 5):
    model_RF = RandomForestRegressor(random_state=12345, n_estimators=estim)
    model_RF.fit(features_train, target_train)
    predictions_valid = model_RF.predict(features_valid)
    print(predictions_valid)
    mse = mean_squared_error(target_valid, predictions_valid)
    rmse = mse**0.5

    print('Estimators:', estim)
    print('Rmse:', round(rmse, 2))
    print('Accuracy:', accuracy_score(target_valid, predictions_valid))
    print('Precision:', precision_score(target_valid, predictions_valid))
    print('Recal:', recall_score(target_valid, predictions_valid))
    print()