A Data Mining Framework for Optimal Product Selection in Retail Supermarket Data: The Generalized PROFSET Model

摘要

In recent years, data mining researchers have developed efficient associationrule algorithms for retail market basket analysis. Still, retailers oftencomplain about how to adopt association rules to optimize concrete retailmarketing-mix decisions. It is in this context that, in a previous paper, theauthors have introduced a product selection model called PROFSET. This modelselects the most interesting products from a product assortment based on theircross-selling potential given some retailer defined constraints. However thismodel suffered from an important deficiency: it could not deal effectively withsupermarket data, and no provisions were taken to include retail categorymanagement principles. Therefore, in this paper, the authors present animportant generalization of the existing model in order to make it suitable forsupermarket data as well, and to enable retailers to add category restrictionsto the model. Experiments on real world data obtained from a Belgiansupermarket chain produce very promising results and demonstrate theeffectiveness of the generalized PROFSET model.