Abstract
Intermittent demand forecasting has been widely researched in the context of spare parts management. However, it is becoming increasingly relevant to many other areas, such as retailing, where at the very disaggregate level time series may be highly intermittent, but at more aggregate levels are likely to exhibit trends and seasonal patterns. The vast majority of intermittent demand forecasting methods are inappropriate for producing forecasts with such features. We propose using temporal hierarchies to produce forecasts that demonstrate these traits at the various aggregation levels, effectively informing the resulting intermittent forecasts of these patterns that are identifiable only at higher levels. We conduct an empirical evaluation on real data and demonstrate statistically significant gains for both point and quantile forecasts.
Highlights
Intermittent demand forecasting is a challenging problem that relates to many aspects of supply chain (Bacchetti and Saccani, 2012), retailing (Fildes et al, 2018) and predictive maintenance (Van der Auweraer et al, 2018), among other applications
The vast majority of intermittent demand forecasting methods provide constant forecasts, assuming that there are no trend or seasonal components in the data. Aggregating these forecasts to larger time buckets results in constant forecasts, even though the data start to exhibit such components. To overcome this dissonance we propose using temporal hierarchies to adjust disaggregate intermittent forecasts to account for identified components at higher aggregation views
We discuss the necessary considerations to forecast with temporal hierarchies on intermittent time series
Summary
Intermittent demand forecasting is a challenging problem that relates to many aspects of supply chain (Bacchetti and Saccani, 2012), retailing (Fildes et al, 2018) and predictive maintenance (Van der Auweraer et al, 2018), among other applications. Croston (1972) proposed a forecasting method to address this complexity by modelling the demand size and interval as two separate entities and dealing with each variability independently. There has been considerable research on alternative forecasting approaches (see, Teunter and Duncan, 2009; Bacchetti and Saccani, 2012; Van der Auweraer et al, 2018). Bootstrapping (for example, Willemain et al, 2004; Syntetos et al, 2015), machine learning and neural networks (for example, Kourentzes, 2013; Nikolopoulos et al, 2016), and model based approaches (Snyder et al, 2012; Svetunkov and Boylan, 2017) have been considered
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.