Research On Points Of Interest Recommendation Integrating Spatio-Temporal Background In Location-Based Social Networks

The next point-of-interest(POI)recommendation is indispensable in enhancing the richness of users’ lives and helping service providers achieve more economic earnings.The next point of interest(POI)recommendation uses the user’s check-in information on location-based social networks to make recommendations.The existing methods based on deep learning are evident in improving the performance of the recommendation model by capturing users’ behaviour dependences.However,existing Recurrent Neural Networks(RNN)based methods lack sufficient interaction with their contexts while treating POI sequences as a straight serial pipeline.On the other hand,most attention-based methods focus on all POIs in the user check-in sequence,even if some attention weights are small.Aiming at the problems existing in POI recommendation scenarios in location-based social networks,this paper propose a more efficient and high-performance recommendation algorithm to improve user experience.The research content of this paper is summarized as follows:Firstly,existing RNN-based methods lack sufficient interaction with their contexts,and at the same time,ignore the importance of non-consecutive POIs with different degrees for understanding users’ behaviors.This paper proposes a novel Spatio-Temporal model based on mogrifier LSTM and attention network(named STMLA)for next POI recommendation.The STMLA model builds a parallel structure to process the users’ check-in sequences through the mogrifier LSTM and the multi-head attention network,which can achieve better contextual interaction while selectively considering non-consecutive factors with different degrees of significance.Our STMLA algorithm explicitly integrates temporal and spatial information to capture users’ long-term and short-term behaviour dependences,incorporating spatial information to build the Location-Saltant algorithm.Secondly,most attention-based methods focus on the global sequence of POIs.This paper proposes a novel spatio-temporal model based on the position-extended algorithm and gated-deep network(i.e.,ST-PEGD)for next POI recommendation.Specifically,by combining spatio-temporal factors,this paper designs a gated-deep network to capture the long-term behaviour dependence of users by generating auxiliary binary gates.In addition,when capturing the short-term behaviour dependence of users,this paper uses the position-extended algorithm to make the contextual interaction of RNNs more sufficient when performing POI sequence hopping selection.Thirdly,the POI sequence of the target user is input into a straight serial pipeline,which does not fully consider the similarity between the target user’s check-in trajectories.This paper improves based on ST-PEGD and proposes the ST-PEGDS model.Specifically,the multi-head self-attention mechanism enhances the model’s ability to select useful information in the user’s check-in sequence.The similarity between the user’s check-in trajectory sequences is calculated by designing a non-local network to better play the role of time information in capturing the user’s long-and short-term behaviour dependences.Through extensive experiments on several real-world datasets,it is demonstrated that the model proposed in this paper has good performance in the next POI recommendation task.