Table 2 Model parameter settings.
Model type | Models | Parameter setting |
|---|---|---|
Traditional machine learning models | SVM | kernel = ‘linear’, Other parameters select default |
Deep learning models | LSTM | neurons1 = 50, neurons2 = 100, neurons3 = 150, batch_size = 64, epochs = 100, Learning Rate = 0.1, Sliding Window = 10 |
Combination model | ARIMA-CNN-LSTM | filters = 512, kernel_size = 2, strides = 1, 3 layers of neurons = 50, batch_size = 64, epochs = 100, Learning Rate = 0.2, Sliding Window = 10 |
CEEMDAN-CNN-LSTM | filters = 512, kernel_size = 2, strides = 1, neurons = 128, batch_size = 100, epochs = 100, Learning Rate = 0.2, Sliding Window = 10 | |
CEEMDAN-LSTM | neurons1 = 128, neurons2 = 100, epochs = 100, Learning Rate = 0.2, Sliding Window = 10 | |
Optimization algorithm model | 3 layers of neurons = [1,300], Sliding Window = [1,50], Learning Rate = [0.001,0.99], batch_size = [1,300], filters = [1,600], kernel_size = [1,10], strides = [1,5], |
