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

Table 8

Evaluation characteristics of tested models.

Model	Error distribution	Akaike	Schwarz	Log-likelihood
ARCH(5)	Gaussian	−6.946032	−6.909037	317.917
	Student	−6.966257	−6.923978	3181.130
	GED	−6.967443	−6.925164	3181.670
ARCH(7)	Gaussian	−6.970504	−6.922940	3184.065
	Student	−6.984941	−6.932092	3191.641
	GED	−6.985553	−6.932704	3191.919
GARCH(1,1)	Gaussian	−7.029560	−7.008420	3205.964
	Student	−7.032833	−7.006409	3208.456
	GED	−7.034504	−7.008779	3209.216
EGARCH(1,1,1)	Gaussian	−7.025497	−6.999073	3205.114
	Student	−7.028507	−6.996797	3207.485
	GED	−7.030426	−6.998717	3208.359
PGARCH(1,1,1)	Gaussian	−7.026622	−6.994912	3206.626
	Student	−7.029612	−6.992618	3208.988
	GED	−7.031268	−6.994274	3209.743
TGARCH(1,1,1)	Gaussian	−7.028705	−7.002281	3206.575
	Student	−7.031598	−6.999888	3208.893
	GED	−7.033244	−7.001535	3209.643

GED: generalized error.