Research On Experts Discovery Method Of Social Network

With the development of the society,the social mission is becoming more and more diversified and complicated.In that case,it's difficult for individual experts to solve problems alone.For the purpose of solving the complex and diversified problem,different experts who form teams,cooperate with each other and share resource are needed.Besides,team members should maintain a good cooperative relationship with others so that the costs can be reduced and the problems can be solved efficiently.How to find teams at lower costs and higher efficiency is important work for solving social problems.The research on the expert finding problem becomes a hot research work in recent years.However,there are some problems for the existing expert finding approaches.First,the existing approaches ignored the relation between nodes in social network structure.Furthermore,since some of the algorithms search the entire network to look for a team,the time overhead as the number of the nodes in social network is very large.To address these concerns,expert finding approach based on social network is studied in this paper.The main work and innovations in this paper include the following aspects.First of all,on the basis of discussion and analysis on basic theory and related technology of research on social network,formal description of expert finding problems on social network,related concepts and definitions of the term are given in this paper.Expert finding problem is divided into three sub-problems,which are individual expert finding problem,expert team finding problem and Top-k expert team finding problem.And the formal descriptions of these sub-problems are given separately.Besides,aiming at above three sub-problems,individual expert finding method,expert team finding method and Top-k expert team finding method are researched.MIREDA algorithm is proposed to find individual experts at minimum cost for individual experts finding problem.The importance of each node in social network is evaluated by adjacent centricity for the double objective optimization problem of expert team finding method,and MCCEDAA algorithm and MCCIREDA algorithm which combine adjacent centricity are proposed.The Top-k EDA algorithm is put forward for finding top k expert team to solve the Top-k expert team finding problem.Meanwhile the main idea and formal description of each algorithm are described,and their time complexity are discussed and analyzed.Finally,the experimental data which meet the requirements of this paper are extracted from DBLP dataset and IMDB dataset,and the data are pretreatment.Experiments are designed and implemented to verify the feasibility and efficiency of each proposed algorithm.The algorithms mentioned above are analyzed and compared from several aspects,which are individual cost in team,cooperation cost in team and team overall cost,the number of team member and running time of algorithm,extendibility of algorithm and so on.The results of experiments indicate the feasibility and correctness of the algorithm.