Research And System Implementation Of Influence Blocking Algorithm In Social Network

In recent years,with the rapid development of Internet technology represented by social networks,people have entered the Internet age.With the emergence and popularity of various social network platforms,researchers can obtain a wealth of behavioral data,which provides the possibility for further research on information propagation,prediction and utilization in social networks.The problem of influence maximization(IM)is an important issue of information propagation in social networks.The problem is to find a group of most influential users and spread information through them to achieve the maximum information propagation.In social networks,users are both receivers and disseminators of information,and the cost of propagation is low.Meanwhile,when information is propagated on the network,there are both positive and negative propagation.Users with positive influence spreading positive information to prevent the spread of negative influence to the maximum extent,this is the problem of negative influence blocking.The blocking of influence spreading is an important issue in the field of network analysis which is theoretically and highly complex.It has a wide range of applications,such as marketing,public opinion monitoring,and rumor blocking.In order to facilitate businesses,enterprises and government departments to monitor,analyze and blocking the influence of negative information in social networks in a timely manner,this paper studies the negative influence blocking maximization(IBM)issues in social networks from two perspectives.One is the negative IBM problem based on determined source,the other is the negative IBM based on uncertain sources,and the corresponding public opinion monitoring and recommendation system has been developed and applied in practice.The main work of this paper is as follows:(1)Propose a negative influence blocking maximization algorithm based on community mining.We first divide the network by negative seeds;then,we merge communities with tighter links by considering the closeness between the communities;based on the result above,we allocate the positive seeds based on the number of negative seeds owned by each community.Finally,the greedy algorithm is used to find suitable positive seeds in each divided community which can achieve the effect of negative IBM.Through experiments on the four social network data sets of Wiki-vote,Epinions,Mobile and NetHEPT in the real world,the results show that our proposed method can achieve more accurate results compared with the classical method.(2)Propose a negative IBM algorithm based on uncertain sources.First,for the influence blocking maximization problem based on determined sources,IBM based on greedy algorithm is proposed.Second,for the IBM based on uncertain sources,sampling method is adopted to transform the problem of(IBM)influence blocking maximization based on uncertain sources to the problem of IBM based on certain sources.Through experiments on three social network data sets of Twitter,Wiki-vote and NetPHY in the real world,the experiments prove that the suppression effect of the proposed algorithm is significantly improved compared with other algorithms.(3)Designed and implemented a network public opinion monitoring and recommendation system based on the analysis of influence spread suppression.The design purpose of the system is clarified firstly,gave the overall framework of the system,and related functional modules is designed and developed,which specifically implemented the main modules such as social network construction,database management,and network user recommendation.Among them,the user recommendation module uses our two algorithms proposed in Chapter 3 and Chapter 4 respectively based on the negative influence blocking maximization of the definite source and the determined source.