Intelligent Soft Computing on Forex: Exchange Rates Forecasting with Hybrid Radial Basis Neural Network

Table 9

Comparison of predictive qualities (out-of-sample predictions, 1-day horizon).

Model	Error distribution	MSE	MAPE
AR(0)-ARCH(5)	Gaussian	0.00001709	0.319356
	Student	0.00001720	0.320744
	GED	0.00001718	0.320459
AR(0)-ARCH(7)	Gaussian	0.00001708	0.319096
	Student	0.00001717	0.320443
	GED	0.00001714	0.320122
AR(0)-GARCH(1,1)	Gaussian	0.00001709	0.319374
	Student	0.00001715	0.320223
	GED	0.00001714	0.320117
AR(0)-EGARCH(1,1,1)	Gaussian	0.00001706	0.318886
	Student	0.00001714	0.320108
	GED	0.00001711	0.319692
AR(0)-PGARCH(1,1,1)	Gaussian	0.00001706	0.318916
	Student	0.00001711	0.319660
	GED	0.00001712	0.319719
AR(0)-TGARCH(1,1,1)	Gaussian	0.00001706	0.318897
	Student	0.00001712	0.319767
	GED	0.00001712	0.319699