Research On Recommendation Algorithm In Location Based Social Network

The popularity and popularity of mobile terminals makes it more and more popular for users to use GPS’s location services. Users can not only online communication, but also by the online development of the specific location of the line of social networking community. The network, which combines geographical location information and social information, is called a geographical location based social network. The geographical location of the social network on the basis of that recommendation function can provide users with real-time information push, with the help of recommendation function, can realize self-efficacy to the recommendations of the geographical position of interest for the user’s friends recommendation and user. Can also be recommended for users in a specific region of the activities within the region, in order to bring the potential of the real business profits, so the recommendation function of social network caused by domestic and foreign researchers of great concern.This research of recommendation algorithm in LBSN is divided into two parts, which are user-recommendation and geographic location recommendation. Based on the location of the social network, the most critical is the establishment of the user model, because it is directly related to the direction of the recommendation. In this paper, the original GPS data will be processed, and then the coordinate values are converted into data sets with semantic meaning, then the clustering of K center points is carried out, and the user model is established. With the established data model, the integrated user experience value and the location popularity, establish a reasonable management relationship between social network and the real world. Through this user model to further complete the user recommended model and the establishment of location recommendation model, complete the user recommendation algorithm and the location of the recommendation algorithm to improve.This paper puts forward a new method to build user model, and based on the content and project collaborative filtering algorithm to improve user similarity, location preference, the output of the user and the location of the proposed algorithm. A series of tests are carried out for the proposed algorithm. The experimental results show that the algorithm proposed in this paper can achieve the accuracy of the user’s friends as well as the location of the proposed algorithm.