Character Analysis And Mining Of User Behavior In Online Social Network

It can be said that there have been the social network ever since human beings began. The study of human behaviors in social networks and interaction principle has been a hot field of academia. However, to the real human beings’ social network, because of the lack of data and experimental conditions, research on social networks has been stuck at the level of qualitative analysis. A variety of online social applications bring large amount of data, vast amounts of user behavior data is recorded. This brings convenience for our better study of human behavior in social networks.In this paper, based on the theory of complex networks, Sina microblogging platform for data sources, we analyzed the behavioral characteristics of microblogging users from the macro level and micro level. User network characteristics generated by user behaviors were analyzed. The microblogging network’s small-world property and scale-free property were verified. Laid the foundation of future research. Weibo users’individual property, feeds posting behavior and reposting behavior were empirically analyzed. The spammers’ behaviors were modeled through comparative analysis. The method that combined incremental SVM and spammers’behavior model were firstly proposed. And the effectiveness of this method were verified by experiment.Experimental results show that this method’s accuracy is high, and it can deal with the spammers’dynamic behavior. Based on the SVM’s advantage that it is not sensitive of data dimension, the method can handle large data sets.