Intention oriented itinerary recommendation by bridging physical trajectories and online social networks

Abstract

Compared with traditional itinerary planning, intention oriented itinerary recommendation can provide more flexible activity planning without the user pre-determined destinations and is specially helpful for those strangers in unfamiliar environment. Rank and classification of points of interest (POI) from location based social networks (LBSN) are used to indicate different user intentions. Mining on physical trajectories of vehicles can provide exact civil traffic information for path planning. In this paper, a POI category-based itinerary recommendation framework combining physical trajectories with LBSN is proposed. Specifically, a Voronoi graph based GPS trajectory analysis method is proposed to build traffic information networks, and an ant colony algorithm for multi-object optimization is also implemented to find the most appropriate itineraries. We conduct experiments on datasets from FourSquare and Geo-Life project. A test on satisfaction of recommended items is also performed. Results show that the satisfaction reaches 80% in average.