Research On Point-of-Interest Recommendation Algorithm Based On Location Social Network

Point-of-Interest(POI)recommendations help users find new locations that match their preferences and play an important role in the personalized services of Location-Based Social Networks(LBSN).The behavior of the user visiting the place is that the behavior in the real world is not as geographically restricted as browsing the webpage in the virtual world,and the user's check-in data in the social network is very sparse,which makes the recommendation performance of the interest point be improved.Severe challenges,while the sparseness of the data creates difficulties for the interrelationship between computing locations.In response to these problems,this paper proposes two methods of interest point recommendation from the perspective of random walk.Firstly,this paper proposes a point-of-interest recommendation algorithm based on geographic factors and random walks for the influence of geographic location information and user check-in behavior in LBSN recommendation,and the difficulty in calculating the relationship between computing locations due to data sparsity.By extracting the implicit features of the user and the location by analyzing the location of the location and the user's check-in behavior,the user's preference for the location is predicted to recommend the point of interest to the user.The recommendation algorithm uses the random walk model to obtain the location relevance degree,and calculates the location correlation coefficient through the relationship between the locations to filter the location relevance degree to extract the location implicit feature,and the user implicit feature vector and location extracted from the check-in data.The implicit feature vector point multiply predicts the user's preference for the location,and the location is recommended to the user according to the user's ranking of the location preference.Secondly,in order to obtain the relationship between the locations more accurately in the sparse LBSN dataset and improve the recommendation performance,a point-of-interest recommendation algorithm based on location implicit features is proposed.The algorithm obtains the relationship between the locations from the two perspectives of random walk and association rules.According to the degree of association between the locations generated by random walks and the association rules,the location confidence,location interest,and location correlation coefficient The location implicit feature is extracted,and the interest point is recommended to the user through the user implicit feature and the location implicit feature.Finally,the two interest point recommendation methods proposed in this paper are tested on the two real data sets Foursquare and Yelp,and compared with two classic interest point recommendation algorithms.