Table 3 Parameter setting of various models.

CC-BP

#1

net.trainParam.goal = 0.0001; net.trainParam.lr = 0.001; net.trainParam.epochs = 500

CC-ELM

#2

For input layer, number of neuron nodes = 3,

For hidden layer, number of neurons = 1,

For output layer, number of neuron nodes = 30

CC-LSTM

#3

Number of nodes in hidden layer 2 = 18; number of nodes in hidden layer 1 = 15

CC-Transformer

#4

Sequence_length = 10, batch_size = 64, feature_size = 250, num_layers = 1, nhead = 10, num_epochs = 100

CC-Informer

#5

Features = MS, seq_len = 384, label_len = 192, pred_len = 96, enc_in = 8, dec_in = 8, c_out = 8, d_model = 512, n_heads = 8, learning_rate = 0.0001, loss = mse

CC-XGBoost

#6

max_depth = 4; learning_rate = 0.05

CC-BiGRU

#7

For hidden layer 2, number of nodes = 20; For hidden layer 1, number of nodes = 10

CC-GBDT

#8

n_estimators = 10; learning_rate = 0.001

CC-GBDT-BiGRU

#9

n_estimators = 10; learning_rate = 0.001; For hidden layer 2, number of nodes = 20; For hidden layer 1, number of nodes = 10;\(\rho =0.3\)

CC-TF-BiGRU

#10

n_estimators = 10; learning_rate = 0.001; For hidden layer 2, number of nodes = 20; For hidden layer 1, number of nodes = 10