A New Hybrid Forecasting Model Based on SW-LSTM and Wavelet Packet Decomposition: A Case Study of Oil Futures Prices

Table 9

Prediction performance evaluation of hybrid forecasting models.

Errors	MAE	RMSE	MAPE	SMAPE	TIC
Index	WTI
CEEMDAN-LSTM	1.2017	2.1849	1.7099	1.8371	0.0167
VMD-LSTM	1.3158	2.5268	2.2632	2.4196	0.0192
ST-GRU	1.1701	2.2165	1.6895	1.9726	0.0151
WPD-SW-LSTM	0.8283	1.8493	1.4329	1.3143	0.0130
Index	Brent
CEEMDAN-LSTM	0.9782	2.3274	1.3412	1.5177	0.0228
VMD-LSTM	1.2814	2.5638	1.4101	1.7065	0.0239
ST-GRU	1.1267	2.2680	1.2826	1.4371	0.0215
WPD-SW-LSTM	0.6756	1.8180	0.9579	1.2339	0.0191
Index	RBOB
CEEMDAN-LSTM	0.0269	0.0381	1.6774	1.5205	0.0168
VMD-LSTM	0.0368	0.0558	1.9783	1.8152	0.0175
ST-GRU	0.0326	0.0328	1.6186	1.4322	0.0116
WPD-SW-LSTM	0.0122	0.0302	1.0350	1.0012	0.0085
Index	Heating oil
CEEMDAN-LSTM	0.0376	0.0805	1.2376	1.5278	0.0183
VMD-LSTM	0.0497	0.1014	1.8509	2.0125	0.0252
ST-GRU	0.0408	0.0732	1.0338	1.2957	0.0171
WPD-SW-LSTM	0.0212	0.0642	0.4775	0.6298	0.0143