Estimation of dominant features of commodities based on shopping behavior analysis

Abstract

It is necessary to provide suitable assistance to each consumer in shopping to choose preferable commodities. Each consumer does shopping with checking dominant features of the commodities according to his own criteria [1]. For example, "I want a cloth of a good material", "I want a T-shirt in cool color", and so on. We have developed an experimental shopping space equipped with ubiquitous sensors such as cameras and RFID-tag readers. In our experiment, each subject freely walked around the shelves to find the preferable T-shirts. Our system observed typically the time of three actions, "Look at", "Touch" and "Take" a T-shirt. In this study, we have tried to estimate the dominant features with each consumer through suggest the approach to recommend information in consideration of personal dominant features from observation and analysis of shopping behavior to perform suitable assistance. © Springer-Verlag Berlin Heidelberg 2013.