Comments on the attribution of an intermittent demand estimator

Syntetos, Aris A. and Boylan, John (2006) Comments on the attribution of an intermittent demand estimator. International Journal of Forecasting, 22 (1). p. 201. ISSN 0169-2070

Abstract

Gardner and Koehler (2005) recently commented on a paper by Willemain, Smart, and Schwarz (2004) pointing out, amongst other issues, some recent advancements on Croston's method (1972) that were not considered. In this commentary the authors misattributed a bias-corrected version of Croston's method to Eaves and Kingsman (2004). The method adjusts Croston's estimates by using a multiplicative factor of (1 − α / 2), where α is the demand interval smoothing constant. Eaves and Kingsman (2004) were the first authors to comment on this method. However, it was originally presented by Syntetos and Boylan (1998); it was subsequently included in the doctoral thesis of Syntetos (2001) and recently published by Syntetos and Boylan (2005). Croston's bias was identified in an earlier work ( Syntetos & Boylan, 2001). We do accept that any error may be explained by the sequence of our publications over the years, and we are indebted to Professor Gardner for giving the correct attribution in his address to the ISF in June 2005. We should be grateful if other researchers would also give the correct attribution of the identification of Croston's bias (Syntetos & Boylan, 2001) and the new bias-corrected method (Syntetos & Boylan, 2005).