Assortment planning using data mining algorithms

Abstract

Assortment optimization is not just selecting the best products according to the sales performance under a certain category, but also an execution method to apply retailers commercial strategy into market considering all strategies which retailer want to play. Regarding millions of data saved in databases and explosive growth of data leads to a situation in which it is increasingly difficult for retailers to understand the right information. To cope with this problem we are planning to use association algorithms to put in place data mining in product selection. It should also be considered that selecting best and suitable products for assortment of retailer need not only sophisticated algorithms to take decisions but also business perspective to embed into decision system. In this study, we approach the assortment selection problem, by improving the PROFSET model and GENERALIZED PROFSET model, which is based on a microeconomic framework. We improved the basic model by introducing additional method of profit allocation over frequent item sets, constraints about categories and sold quantities. Finally we empirically test our model with sample retailer data. While doing this we will also take into consideration the retail industry characteristics and consumer and customer perceptions.