MLP_Tester.py

"""
Module: MLP_Tester.py
Use: Reads in the pickled MLP and testing sets from MLP_Creator, determines MAPE, and plots the results
Last Edited: Akrit Sinha, 06-28-2019
"""
# Packages
import pickle
import numpy as np
import matplotlib.pyplot as plt

# Import pickled testing sets
X_test = pickle.load(open('X_test', 'rb'))
y_test = pickle.load(open('y_test', 'rb'))


# Defines MAPE function for accuracy
def mean_absolute_percentage_error(y_true, y_pred):
    y_true, y_pred = np.array(y_true), np.array(y_pred)
    return np.mean(np.abs((y_true - y_pred) / y_true)) * 100


# Initiates 500 runs and determines average MAPE score
if __name__ == "__main__":
    load_mlp = pickle.load(open('MLP', 'rb'))
    predictions = load_mlp.predict(X_test)
    print(100 - mean_absolute_percentage_error(y_test.values.ravel(), predictions))
    plt.plot(y_test.values.ravel(), 'k')
    plt.plot(predictions, 'y--')
    plt.show()