The Online Social Network User Behavior Analysis Based On The Network Structure

Social networks have permeated into everywhere of our society, such as online social networks, research networks, and traffic networks, etc. With the development of information science, many scholars pay attention to the social behavior. The structured data often represent by complex networks, the node is the research object and the edge is the interaction or relationship between objects. Not only from a scientific perspective but also for commercial or strategic motivations, social network analysis has become a very important research method. Social network analysis method involves a wide range of disciplines and fields, such as data mining, data visualization, statistical analysis, information dissemination, etc.This paper selects the two representative platforms of the online social network, Sina and QQ. Network topology analysis and user behavior research is carried on the platform. Finally according to the actual data, we discuss the interaction and influence of structure and behavior of online social networks.In this paper, the first chapter is introduction, the research background, significance and the current situation of the social networks are included. The second part introduces the basic concept and methods of the social networks, including the basic theory, definition of social networks, the characteristics of the network structure and the analysis of user behavior. The third chapter and the fifth are the core parts of the paper. The third chapter involves the social network structure and user behavior, through the data of sina and QQ. We construct the network by Pajek software to learn the network structural properties and behavior.In the fourth part the node centrality ranking algorithm is proposed to calculate the significant nodes in social networks. Based on the ranking algorithm of nodes, this paper proposes the edge ranking algorithm, and verifies the correctness of the algorithm.In the fifth chapter, we analyze the relationship between the structure and the behavior and then draw the conclusion that the degree of the network determines the user’s behavior to some extent. The last part of the paper is the summary and outlook, concluding the entire contents of the paper. At the same time, the shortcoming of this paper and the further resolved problem of signed network are shown.In this paper, the data is visualized in the form of network, and the relevant models are established to describe the user’s behavior. We analyze the collected data by SPSS, Excel and other statistical tools and explore the inherent relationship between the variables from multiple perspectives. Finally, we use hypothesis testing to explore the relationship between the variables. This paper includes the following innovations: The first, node centrality and edge centrality algorithm based on network structure is proposed, and the correctness of the ER algorithm is discussed. The second is to obtain the statistical data of Weibo user behavior and analyse it. The third, we use the hypothesis testing to explore the relationship between the various indicators.