Retail recommendation system

Abstract

Market basket analysis (also known as association-rule mining) in modern retailing has been a useful method in discovering customer purchasing patterns by extracting associations or co-occurrences from stores’ transaction databases which are readily available recently because of the growth in technology. This information obtained from the analysis can be used in forming marketing, sales, service, and operation strategies. The existing methods, however, fail to discover important purchasing patterns in a multi-store environment, because of an implicit assumption that products under consideration are on shelf all the time. In this study, i propose a new method to overcome this weakness. My empirical evaluation shows that the proposed recommendation system is computationally efficient, and it has advantage over the traditional method when stores are diverse in size, product mix changes rapidly over time, and larger numbers of stores and periods are considered.