Forecasting of Sporadic Demand Patterns with Seasonality and Trend Components: An Empirical Comparison between Holt-Winters and (S)ARIMA Methods

Table 2

Selected (S)ARIMA models for 5 and 12 time periods ahead.

Series	Group	5 time periods ahead	12 time periods ahead
s1	1	(3,1,1) $\times$ (2,0,2)4	(2,1,3) $\times$ (2,0,2)4
s2		(2,1,2) $\times$ (3,0,1)4	(1,1,3) $\times$ (2,0,2)4
s3		(3,1,1) $\times$ (3,0,2)4	(3,1,2) $\times$ (2,0,2)4
s4		(1,1,2) $\times$ (2,0,3)4	(1,1,3) $\times$ (2,0,2)4
s5		(2,1,2) $\times$ (2,0,2)4	(3,1,2) $\times$ (2,0,1)4
s6		(1,1,2) $\times$ (2,0,2)4	(3,1,2) $\times$ (2,0,1)4
s7		(2,1,2) $\times$ (3,0,1)4	(1,1,1) $\times$ (3,0,1)4
s8	2	(3,1,2) $\times$ (2,0,2)4	(1,1,1) $\times$ (3,0,1)4
s9		(2,1,2) $\times$ (3,0,1)4	(2,1,2) $\times$ (2,0,1)4
s10		(1,1,2) $\times$ (3,0,1)4	(2,1,3) $\times$ (2,0,1)4
s11		(2,1,2) $\times$ (3,0,3)4	(1,1,1) $\times$ (2,0,2)4
s12		(2,1,2) $\times$ (2,0,1)4	(3,1,2) $\times$ (2,0,1)4