Home > Research > Publications & Outputs > Generalised estimators for seasonal forecasting...
View graph of relations

Generalised estimators for seasonal forecasting by combining grouping with shrinkage approaches

Research output: Contribution to Journal/MagazineJournal articlepeer-review

<mark>Journal publication date</mark>1/03/2013
<mark>Journal</mark>Journal of Forecasting
Issue number2
Number of pages14
Pages (from-to)137-150
Publication StatusPublished
Early online date9/10/11
<mark>Original language</mark>English


In this paper, generalised estimators are proposed to estimate seasonal indices for certain forms of additive and mixed seasonality. The estimators combine one of two group seasonal indices methods—Dalhart's group method and Withycombe's group method—with a shrinkage method in different ways. Optimal shrinkage parameters are derived to maximise the performance of the estimators. Then, the generalised estimators, with the optimal shrinkage parameters, are evaluated based on forecasting accuracy. Moreover, the effects of three factors are examined, namely, the length of data history, variance of random components and the number of series. Finally, a simulation experiment is conducted to support the evaluation