Improvement Of Location Recommendation Algorithm Based On Space-time And Socialized Network Relationship

In recent years, the location-based social network has been developed rapidly. Extracting the user’s spatial positions and social network relevance to establish reasonable location recommended models according to the user’s social activities and users’ check-in data has become the hot topic for study based on the location-based social network recommendation systems.This paper firstly mined the characteristics of Gowalla users’ check-in data in time, space and social network domains. And then put the time, space and users’ social network relationships into the framework of location-based recommendation system, thus, the location can be recommended to users. In addition, the analysis and comparison between the present algorithm and traditional Group Recommendation Algorithm Based on User Collaborative Filter have also been presented in this paper.The experimental results show that the User Location Recommendation Algorithm Based on Space-time and the Location Recommendation Algorithm Based on Space-time and Social Network Relationship proposed in this paper have better accuracy and recall rate than the User Collaborative Filtering Recommendation Algorithm Under Time Slice Division.In the following section, this paper summarizes the steps of group division, user preference fusion, group recommendation algorithm and effectiveness evaluation and proposes a Group Location Recommendation Algorithm Based on User Clustering which gets good recommended results,after experimenting on the Gowalla dataset. Finally, this paper concludes the research contents of this paper, analyzed and prospected the difficulties and hot issues that need further studies.