Research On The Community Structure And The Robustness Of Social Networks

With the popularity of digital mobile terminals and the convenience of the Internet,social network data are growing explosively.The explosion of data leads to the advent of a new era,and this era is called Big Data.Big Data is a new and cross-disciplinary science,and the structural analysis of social networks is now becoming the cornerstone to big data science.The analysis of the structural characteristics of social networks is of great significance to both public sentiment analysis and mining potential value of social networks.This paper mainly investigates the community structure and the robustness structure of signed social networks and the detailed works are as follows:1)Introduces the backgrounds related to complex networks,which mainly includes the con-cept and notations of networks,the basic network properties like community structure prop-erty and network robustness.Reviews the current literature on the community structure of complex networks,summarizes the related works on the robustness structure of complex net-work,illustrates the fundamentals of evolutionary optimization,introduces the basic ideas and principles of the standard particle swarm optimization algorithm;2)With regards to the community mining problem of signed social networks,this paper im-proves the multi-objective particle swarm optimization based community structure mining algorithm existed in the literature,and an improved algorithm for network community de-tection is proposed.In the algorithm,a new multi-objective community mining model is put forward.In order to enhance population diversity,this paper improves the subproblem up-date strategy.For the sake of better intelligent decision making,this paper proposes the new multi-criteria decision-making method.The improved method is validated on a large number of synthetic and real-world signed social networks and its high effectiveness is proved;3)This paper does research on the robustness of social networks.First,it systematically investigates the robustness of a social network to nodes loss under three different attack strategies.Then,it studies the robustness of a social network to community perturbation-s under three different attack strategies.During the analysis of community robustness,it proposes two different indices to measure the importance of a community.Simulation ex-periments on real-world network data show that social networks are robust to random node or community attacks,but fragile to target ones.