An Outdoor Recommendation System based on User Location History

Recommendation systems are now widely used in online shopping sites, but so far there have been few attempts of applying them to real-world shopping. In this paper we propose a novel real-world recommendation system, which recommends shops to users based on their individual preferences and needs, estimated by analyzing their past location history acquired using GPS. The system automatically figures out each user's frequently visited shops using a custom estimation algorithm, and makes recommendations by using the shops as input to the item-based collaborative filtering algorithm. Furthermore, to provide more timely recommendations, our system takes into account the user's usual shopping routes, and the ease of access from the user's current location to each shop. We have conducted an evaluation test using a prototype of our system at Daikanyama, Tokyo, and the results show great promise in the system's ability to make accurate recommendations.