Analyzing And Modeling Of The Microblog Topic Propagation Characteristics

Recently,microblog platform becomes a popular source of all kinds of information.It provides a convenient way for people sharing ideas,opinions,photos and videos.On the one hand,microblog platform has improved the way people communicate with each other.On the other hand,it also make rumors,defamations and vituperations spreading extremely fast and abroad which brings bad influences on the individuals and the society.How to detect the bad information at the early stage and how to effectively control the spreads of them are the problems before us.To solve those problems,two basic problems are present:1)how the information spread in microblog platform;2)what are the common features of information spreading among users.These are what this article tries to address.Concentrated on the user behavior,this paper analyzed the characteristics of microblog system and the topic propagation,proposed a user behavior decision-model based on the multiple features to predict the topic trending.First,based on the study of user behavior and topic propagation on real microblog data,a Petri-Net model is established.The construction of social network and the process of topic propagation are two critical sections that lie in the way of microblog system simulation-model.Further analysis also shows that the structure of social network has a great impact on the process of topic propagation.Following the microblog system simulation,further study on social network is going on to understand the importance of characteristics of the network structure and correlation among these characteristics.Compared to the real microblog data,the existing models have many disadvantages on simulating user relation network.With these efforts,a new network model,based on the Forest Fire model,is proposed.The validation of the advanced model via emulation experiments followed later.Then,a user behavior decision-model is established and analyzed.The decision-model is based on multiple features which can be classified into three aspects user behavior,topic factor and user relation strength.We also analyzed the impact of the changing network structure on topic propagation.The conclusion of the whole study and the suggestions of further study are present at the end.