Model

Parameter settings

SVR

kernel = 'rbf', C = 100, gamma = 0.001

BP

string = {'logsig', 'purelin'}, func_str = 'trainlm', the number of neurons = [3, 2], learning rate = 0.001, epochs = 500

ELM

C = 10 (The inverse of the regular coefficient), the number of neurons = 5

LSTM

the number of neurons = 128, learning rate = 0.0001, batch_size = 64, epochs = 500

GAN

learning rate = 0.0001, batch_size = 64, epochs = 500

WGAN-GP

learning rate = 0.0001, batch_size = 64, epochs = 500