Intelligent Recommender System Using Shopper's Path and Purchase Analysis

Shoppers having a predefined shopping list usually follow the shortest path through a supermarket or store in order to make their purchases. This paper aims to study customer behaviour of such shoppers with respect to two aspects: (1) the path followed through the store to make those purchases. (2) the average path length to make those purchases. The paper also proposes a methodology consisting of a number of stages such as data cleaning, path generation, clustering and classification, association rule generation, determination of attraction values that correspond to the appeal of each product or location and calculation of metrics such as average path length of paths taken by customers, average number of purchases per bill for trend analysis. This research can be used to recommend certain changes, in terms of the store layout, in order to increase the attraction and sale ability of the various products in different locations in the store as an effect of modified store layout.