ml_timeseries_example.py

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt


def model(x, a2, a3):
    return a2 * x + a3 * x ** 2


def generate_training_data(x, a2_range, a3_range):
    data = []
    a2_draw = []
    a3_draw = []
    a2_samples = np.random.uniform(a2_range[0], a2_range[1], 100)
    a3_samples = np.random.uniform(a3_range[0], a3_range[1], 100)

    for a2 in a2_samples:
        for a3 in a3_samples:
            y = model(x, a2, a3)
            a2_draw.append(a2)
            a3_draw.append(a3)
            data.append(y)

    return np.array([a2_draw, a3_draw]), np.array(data)


def generate_ml_model(in_params, out_timeseries):
    model = tf.keras.Sequential([
        tf.keras.layers.Dense(in_params.shape[0], activation='relu'),
        tf.keras.layers.Dense(64, activation='relu'),
        tf.keras.layers.Dense(out_timeseries.shape[1])
    ])
    model.compile(optimizer=tf.keras.optimizers.Adam(0.001),
                  loss='mse',
                  metrics=['mae', 'mse'])
    history = model.fit(in_params.T, out_timeseries, epochs=20, verbose=1)
    # plot loss from history
    plt.plot(history.history['loss'])
    plt.title('model loss')
    plt.ylabel('loss')
    plt.xlabel('epoch')
    plt.savefig('loss.png')

    return model


if __name__ == '__main__':
    x = np.linspace(0, 10, 100)
    a2_range = (-5, 5)
    a3_range = (-5, 5)
    in_data, out_data = generate_training_data(x, a2_range, a3_range)
    ml_model = generate_ml_model(in_data, out_data)
    model_eval = ml_model.predict(np.array([[0.5, 0.5]]))[0]
    plt.figure()
    plt.plot(x, model_eval, label='ML')
    plt.plot(x, model(x, 0.5, 0.5), label='True', ls='--')
    plt.legend()
    plt.show()