Offline Shop Recommendation based On Online Shopping History

Most existing POI recommender systems is facing the problem of sparsely of data. In this paper, an O2O (Online to Offline) recommendation method is proposed, which can take advantage of users' online shopping history data. User behavior is converted to user preference, modeled as user and interest-term matrix. Based on the model, the result list of offline shop is recommended by O2O recommendation algorithm. The algorithm is implemented, and experiments on the real dataset show the feasibility of the method.