Forecasting Crude Oil Price with Multiscale Denoising Ensemble Model

Table 2

In-sample forecasting accuracy comparison of wavelet models for WTI market.

Models	MSE×10^−3	MSE×10^−3,data	MSE×10^−3,noises
Haar	1.4273	1.4189	1.3880
Db(2)	1.3988	1.3890	1.3890
Db(3)	1.4094	1.4140	1.3791*
Db(4)	1.4152	1.3886	1.4083
Db(5)	1.4141	1.3934	1.3987
Dmey	1.3954	1.4166	1.3573*
Coiflet(1)	1.4263	1.4228	1.3844*
Coiflet(2)	1.3907	1.3865	1.3849
Coiflet(3)	1.7108	1.4823	1.6104
Coiflet(4)	1.4031	1.3982	1.3863
Coiflet(5)	1.3570*	1.3677*	1.3688*
Symlet(1)	1.4273	1.4189	1.3880
Symlet(2)	1.4318	1.4233	1.3884
Symlet(3)	1.4094	1.4140	1.3791*
Symlet(4)	1.4037	1.4106	1.3744*
Symlet(5)	1.4510	1.4305	1.4053
Bior(1, 1)	1.4273	1.4189	1.3880
Bior(2, 2)	1.4130	1.4105	1.3863
Bior(3, 1)	1.7564	2.3318	1.8394
Bior(3, 9)	1.4373	1.4133	1.3982
Rbior(1, 1)	1.4273	1.4189	1.3880
Rbior(2, 2)	1.4647	1.4141	1.4294
Rbior(3, 1)	1.4359	1.4248	1.4521
Rbior(3, 9)	1.3790*	1.3816*	1.3789*

∗The test case with superior performance than that of Random Walk model.