Using Association Rules for Product Assortment Decisions in Automated Convenience Stores

Association rules is a recent data mining technique to discover affinities, in large transaction databases, between items frequently purchased together. It has been claimed that the discovery of frequent sets of items is well suited for applications of market basket analysis to discover regularities in the purchase behaviour of customers. In this study, we integrate the discovery of frequent itemsets from association rules with an integer programmming model for product selection (PROFSET). The model combines both quantitative and qualitative (domain knowledge) criteria. Sales transaction data obtained from a fully-automated convenience store are used to demonstrate the effectiveness of the model against a heuristic for product selection based on product-specific profitability. We show that with the use of frequent itemsets we are able to identify the cross-sales potential of product items and use this information for better product selection. Furthermore, we demonstrate that the impact of product-mix decisions on overall assortment profitability can be evaluated by means of sensitivity analysis.