Table 2 Parameter settings of the compared models

GRNN

MSE = 0.001 to start the evolution, and the maximum hidden layer is 10

LSTM

250 epochs of training, Gradient t, Threshold = 1, Learning Rate = 0.005, hidden units = 200

RNN

Lag = 1:1; horizon = 50; MiniBatchSize = 64; MaxEpochs = 600; learningrate = 0.00611

SVRPSO

eps = 1e-3; v = 5; C1 = 1.5; C2 = 1.5;

SVRALO

Agents_no = 10; dim = 10; Max_iteration = 100; lb = 0.01, ub = 100

CNN

Lag = 1:1; horizon = 50; MiniBatchSize = 64; MaxEpochs = 600; learningrate = 0.00611