Personalized Itinerary Recommendation Based on User-Shared Trails

摘要: Itinerary is a kind of abstraction human life. However, with substantial effort involved, it complicated for someone to plan an itinerary. Therefore, personalized and reasonable itinerary recommendation nontrivial. With the development social networks, people have shared their trails like photos, check-ins text comments on locations. Historical itineraries can be extracted from these make possible recommend new ones. Existing studies mainly focus isolated They firstly choose eligible locations then construct filtered various contexts. connections between been ignored. In this paper, we propose concept transfer. We predict user preferences based actions transfers link itineraries. evaluate our method foursquare check-in dataset achieved promising results.