Table 12.

Mean of the root mean square errors for model forecasting ($M=1,5,$ and 10 steps-ahead), computed for each of the four models, ARIMA, SSA, robust SSA based on the $L_1$ norm, and robust SSA based on the Huber function, for the simulated data, based on 100 runs, using $L=24$ and $r=5$.

M	% of Cont.	Shift	Method
			ARIMA	SSA	RLSSA
M = 1	0%	-	1.685	0.125	0.245
	5%	$y_i+2$	0.843	0.475	0.330
	10%	$y_i+2$	0.793	0.596	0.426
	5%	$y_i\times 5$	3.960	8.461	0.358
	10%	$y_i\times 5$	4.359	9.692	0.652
M = 5	0%	-	1.631	0.122	0.222
	5%	$y_i+2$	0.984	0.475	0.307
	10%	$y_i+2$	0.768	0.586	0.413
	5%	$y_i\times 5$	3.789	538.447	0.323
	10%	$y_i\times 5$	3.853	17.670	0.720
M = 10	0%	-	1.381	0.127	0.244
	5%	$y_i+2$	1.320	0.601	0.358
	10%	$y_i+2$	1.148	0.698	0.474
	5%	$y_i\times 5$	3.486	22.695 *	4.015
	10%	$y_i\times 5$	3.694	622.783	2.320

* 10% trimed mean. The mean value is $1.566\times {10}^6$.