POI Recommendation with Interactive Behaviors and User Preference Dynamics Embedding

Recently, Location-based Social Networks (LBSNs) have been widely used in people's daily life. This generates the demand for effective and efficient POI recommendation. Existing research focused on improving the POI recommendation performance by some extra information such as social links, geographical distance and others, but they failed to integrate those complex influenced factors together. In this paper, we propose a POI recommendation framework with interactive behaviors and user preference dynamic embedding to consider the social and geographical information with interactive behavior embedding, along with users' preference dynamics embedding, respectively. To combine the interactive behavior and users' preference dynamics behavior, we use a fusion parameter to decide the contribution of interactive behaviors embedding and user preferences embedding. Extensive experiments on two real-world datasets demonstrate the effectiveness of our proposed model.