Location-based Social Network User Behavior Analysis And Research

With the development of the information age, location-based services have become a research objective in many fields, and social network has become widely used as a platform for users to share and exchange information. The combination of location-based services and social network provides an important source of data to analyze user behavior. It will help to improve the quality of personalized service to users. Research on similarity between users has important significance in analyzing individual user history behavior and forecasting their future behavior. However, due to the diversity of the data, analysis of user similarity and prediction of user behavior have become the industry’s hot and difficult issue.In this paper, taking into account the diversity of the characteristics of user data, we analyzed and studied the user behavior of location-based social network based on the foursquare data. Firstly, according to the structured data, we propose a potential user characteristics data mining algorithm based on matrix decomposition. While according to unstructured data, we adopt an analysis method which based on topical modal and merge these two kinds of information together to analyze the similarity between users simultaneously. Secondly, we analyze the common characteristics between the language model and location-based social network, and propose a behavior prediction algorithm based on language model. Combining with the similarity between users and other characteristics, we predict the user’s next access location using decision tree. Finally, experimental analysis and verification shows that the similarity analysis based on the fusion of two kinds of information has a good effect, and it has a certain role in promoting for the prediction of the behavior based on language model.