A Sequential Multi-type POI Recommendation Service Based on Location

Abstract

Nowadays locaiton-based services (LBSs) are becoming more and more popular. However, all the existing service can only support the single POI recommendation, which is the typical LBS. A user sends his current location and the single type of POIs (Points of Interest) to the server. Then the server will return proper POIs to the user. In people's life, we have the need of multiple types service. That is, what the user wants is some combinations of different POIs rather than single POIs. For example, one restaurant, one cafe and one cinema. In addition, the visit time of these POIs are different. So the order of each POI in one combination is up to the user. More precisely, what the user need is sequential multi-type POI recommendation. To support this service, we propose a new framework where the interaction between the user and the server is changed. This new service is totally different from the typical one. Therefore, a new algorithm to solve this challenging problem is proposed too. It adopts a recursion idea. At the last, some simulations are conducted to testify the effectiveness of the algorithm based on real data from the Internet.