Estimation of dominant attributes of product for each customer through behavior observation of shopping

Abstract

Our purpose is to make a personalized shopping support system in a retail store. In this study, we estimated dominant attributes for each customer through behavior observation of shopping to collecting decision-making data of them with various kinds of products. The dominant attributes are estimated by conjoint analysis of the product attributes and the degree of interest in the product estimated from customer's behavior. In the experiment with a trial retail store, we achieved success about estimations of 3 of 4 customers. The products recommended by our system also were shown to be better suit for customers. © Springer-Verlag Berlin Heidelberg 2013.