Study On Recommendation Method Based On Location Service Social Network

With the rapid development of social networks, as well as the 3G, 4G and WiFi technology, makes, including mobile phones, tablets, etc, with the function of positioning equipment quantity big increase. So that people can get their own position at any time and any place, with the integration of the location of the information and the traditional social network, the combination of online activities and information on the offline, resulting in the location-based social network. With location information of user data on the Internet is being collected by social networking sites, how to use these data to provide better services to users has become the focus of the current study. At present, there is a very wide range of applications is the location of the recommendation, and the location of the recommendation can be divided into local locations recommended and remote locations recommended.In this thesis, we mainly study the remote location of the recommendations, and proposed a remote location recommendation method and propose a multi relational model, the multi relational model can be in a simple framework for effective combination of the relationship between user and location, the relationship between the user and the user and the relationship between location and location and use restart random walk algorithm to analyze the model. Through using real data from Flickr, this thesis comprehensively evaluates the performance of the method proposed in this thesis. The experimental results show that this method both on personal recommendations and group recommendation has better performance.