Point Of Interest Recommendation Algorithm Based On Social And Geographic Information

In recent years,with the rapid development of smart devices,social networks and location acquisition technologies,location-based social networks(LBSNs)have emerged,such as Foursquare,Gowall and WeChat etc.Location-based social networking services have become essential tools for people to get around.These location-based social networking services combine users' online and offline activities.Users can visit their point-of-interest in real life,check-in actibities online,communicate with social friends online and share their geographical location and check-in experience.The point-of-interest recommendation system is designed to recommend places of interest to users to provide a better user experience and increase business profits for merchants.Because of the point-of-interest recommendation system has such great value to users and businesses,the research of the point-of-interest recommendation system has received extensive attention from academia and industry.Integrating multiple influencing factors to improve the user experience of the point-of-interest recommendation system is a technical implementation approach,but there are also some challenges.Because the recommendation system of point-of-interest is affected by context information such as geographical factors,time factors and social factors,the data in the recommendation system of the point-of-interest recommendation system is very sparse,and it is not easy to accurately recommend the interest points.Based on the existing research work,this paper focuses on the shortcomings of the existing point-of-interest recommendation model,and conducts research on the recommendation technology of point-of-interest based on social and geographic information.By using the user's history to check in data,social relationships and geographic location,this paper explores the data.Users' preferences are mined to achieve accurate recommendations point-of-interest to users.Finally,the validity and reliability of the theoretical research on the recommendation of point-of-interest are verified by the existing published data sets.The main work and innovation of this paper are reflected in the following three aspects:1.Aiming at the cold start problem in the point-of-interest recommendation system,this paper uses the Dirichlet distributed topic model to mine users' topic interest and analyze users'similarity;The Louvain community algorithm and the user's check-in data are used to calculate the user similarity;uses the geographic information to mining users' preference for geographical location;Finally,similarity and geographical information are integrated into a unified model,and users are recommended for point-of-interest.2.This paper proposes a point-of-interest recommendation model that integrates user,social and geographic information.The model studies the interaction between users through the asymmetric user influence and the user global impact factor generated by the PageRank algorithm.An algorithm for calculating the similarity between users based on the living distance of two social users and the common friends of users is designed;Power law distribution is used to mine the influence of distance between the point-of-interests on user's Check-in intention;Finally,the improved user-based collaborative filtering,social impact and geographic impact are combined to generate a point-of-interest recommendation model.3.Using the real check-in data set on Foursquare and Gowall,the research theory of the point-of-interest recommendation model based on social and geographic information proposed in this paper is experimented and analyzed.The results of experimental shows that the recommendation model based on social and geographic information proposed in this paper has a greater imporbement in recommendation effect than the existing point-of-interest recommendation,which indicates that the model is reliable and effective.