Teunter, R H and Sani, B (2006) On the bias of Croston''s forecasting method. Working Paper. The Department of Management Science, Lancaster University.
Abstract
Croston's forecasting method has been shown to be appropriate in dealing with intermittent demand items. The method, however, suffers from a positive bias as shown by Syntetos and Boylan (2001, 2005) who proposed a modification. Unfortunately, the modification ignores the damping effect on the bias of the probability that a demand occurs. This leads to overcompensation and a negative bias, which can in fact be larger than the positive bias of the original method. Levén and Segerstedt (2004) also proposed a modified Croston method, but that suffers from an even more severe bias. Building on the results of Syntetos and Boylan (2001, 2005), we propose a new modification that takes the damping effect into account. A numerical study confirms that it considerably outperforms the existing methods. Moreover, the performance is better over the entire range of relevant parameters, which avoids the need to use different methods depending on the demand categorisation as suggested by Syntetos et al. (2005).