A Hybrid Approach Integrating Multiple ICEEMDANs, WOA, and RVFL Networks for Economic and Financial Time Series Forecasting

Table 2

The settings for the parameters.

Method	Parameters	Description
ICEEMDAN	Nsd = 0.2	Noise standard deviation
	Nr = 100	Number of realizations
	Maxsi = 5000	Maximum number of sifting iterations
LSSVR	Rp = 2^{−10,−9, …, 11,12}	Regularization parameter
	WidRBF = 2^{−10, −9, …, 11,12}	Width of the RBF kernel
BPNN	Nhe = 10	Number of hidden neurons
	Maxte = 1000	Maximum training epochs
	Lr = 0.0001	Learning rate
WOA	Pop = 40	Population size
	Maxgen = 100	Maximum generation
MICEEMDAN-WOA-RVFL	Nsd = [0.01, 0.4]	Noise standard deviation in ICEEMDAN
	Nr = [50, 500]	Number of realizations in ICEEMDAN
	Maxsi = [2000, 8000]	Maximum number of sifting iterations in ICEEMDAN
	Nhe = [5, 30]	Number of hidden neurons in RVFL
	Func = {sigmoid, sine, hardlim, tribas, radbas, sign}	Activation function in RVFL
	Mod = 1: Regularized least square,	Mode in RVFL
	2: Moore–Penrose pseudoinverse
	Lag = [3, 20]	Lag in RVFL
	Bias = {true, false}	Bias in RVFL
	Rand = {1: Gaussian, 2: Uniform}	Random type in RVFL
	Scale = [0.1, 1]	Scale value in RVFL
	ScaleMode = {1: Scale the features for all neurons,	Scale mode in RVFL
	2: Scale the features for each hidden neuron,
	3: Scale the range of the randomization for uniform diatribution}