Research On Online Social Network Information Propagation Model Based On PageRank

As Internet technology develops,online social networks is becoming increasingly popu-lar.Online social network service platform,especially represented by weibo,which has large numbers of user groups,formed enormous information transmission networks.Studying in-formation transmission and analysing the information polarity in online social network can effectively monitor the public opinion and block the spread of bad remarks in the social net-work.The research on the mechanism of information propagation in online social networks is to construct a reasonable communication model.The network topology is embedded into complex network,and the communication model is established based on the communication dynamics theory.At the same time,the use of machine learning and other theoretical methods to analyze user behavior can further reveal the law of information dissemination.Most of the current information dissemination studies do not fully consider the topological characteristics of online social networks.In this paper,we explore the influence of network topology on information dissemination.The main contents are as follows:(1)Mining online social network topology and its relationship with information dissemi-nation.Based on the analysis method of complex network,this paper analyzes the basic charac-teristics of online social network structure,and points out the application of network topology in the research of the current online social network information dissemination.(2)Proposing a information propagation model in online social network based on PageR-ank which is called P-SIR.In this model,the PageRank of the nodes in the network is used as the authority of the nodes.Combined with the epidemic model,we fully consider the topo-logical characteristics of online social networks.Using six different networks to simulate the spread of information.The simulation results show that the proposed model is more effective than the traditional model,and the proposed model is more scalable than the traditional model.(3)Weibo sentiment classification research based on the network topology.Geting the Sina weibo data sets through web crawler technology to build the weibo network for analy-sis.The effectiveness of the network is verified by the P-SIR model.Then combined with the graph theory abstract representations of the relationship between the user and the user behavior and simplify the network topology as relationship-behavior.Extraction of weibo information text class features and network topology class features.Using word2vec to represent text word vector.In order to classify the sentiment polarity of weibo texts,four kinds of machine learning methods were used to train different models on the three feature sets.The experimental results show that the Random Forest algorithm has the best classification effect,and it can improve the classification accuracy of the model and the characteristics of the network is more important.