BBS User Behavior Analysis

With the rapid development of Internet and Web2.0, the virtual space of exchange and communication comes into being, for example, Blog, BBS. In BBS, people establish contact through posting and replying to formate a new form of interaction in virtual communities. Because of the different characteristics between virtual communities and real communities, many domestic and international researchers have studied the BBS user behavior.Now the study of BBS user behavior mainly focuses on two aspects. The first aspect is macroscopic analysis in BBS user behavior, that is, constructing BBS reply networks to research the small world and scale-free features of BBS reply networks; the second aspect is microcosmic analysis in BBS user behavior, we can master characteristics of the user behavior and the forms of user interaction through analysising microscopic interactions.In this paper, the reply networks are constructed with the data downloaded from SINA BBS. Topology features of BBS reply networks are analyzed based on the complex networks theory. The small world and scale-free features of BBS reply networks are researched. We clear the role of the user in the BBS community and the characteristics of interactive behaviors through quantitativing analysis the BBS user behavior. It has important significance on protecting community health development.The contents in the paper are organized as follows:(1) The BBS reply networks will be built according to the response relationship between users, and the visualization of reply-networks will be showed by using pajek software. Based on the complex networks theory, we analyzing topological structure of BBS reply networks by using clustering coefficient, average path length, degree distribution and the index of the node centrality.(2) We count user postings and user ID to quantify user behavior, in order to find user behavior characteristics.(3) Using the number of total posts、topic posts、the topic post got the replies、replies and sticky posts as classification index, and classify the users into strength type, leader type, response type and browser type. Then we analyze the user’s behavior and interaction characteristics based on the similarity of members and the association strength of members.(4) The subject tree is introduced to build response relationship between users, and the topic depth and the topic breadth are introduced to describe the user interaction on the subject. Through anglicising the theme breadth coefficient W and the theme depth coefficient D to evaluate interactions themes, and in accordance with the interactive features of the user’s interaction divided into five categories, which have an active guidance on BBS network for information management.