Analysis Tool's Design And Actualization Aimed At Academic Social Network

Academic social network is a kind of social network which is made of information like literature,scholars,conferences,journal and citation relations.Many types of academic research can be conducted by means of academic social network,just like expert research,the study of the laws of disciplines and literature ranking prediction.So analyzing the academic social network and extracting the features which users demand has important significance in theory and practice.There are three kinds of analysis tools which can be used:Mendeley,Pajek and GraphMarker.Mendeley can custom user's literature database and arrange essential information of each paper.But Mendeley can't deal with the network which is formed by papers.Pajek can graphically represent the structure of the net,it can also calculate coefficients of the network.But Pajek is lack of the ability to analyse the information of each single node.GraphMarker can dynamically generate the image of citation network and coauthor network,but it doesn't have the ability to calculate features of the network.In view of the fact that there doesn't exist an analysis tool which can be conducted on academic social network,this paper designed and actualized a new kind of analysis tool aimed at academic social network.It mainly achieves the following functions:(1)Data preprocessingExtracting the basic information of each paper,then extracting the citation relations according to citation message.Storing data in a standardized form.(2)The citation network module's design and actualizationExtract basic information of each paper,includes the number of papers,time distribution,proportion of the key literature,papers' citation number and citation relations.Achieve the extraction of relevant literature by use of VSM model.Compile the distribution of subject literature.Construct the citation network according to citation relations.Implement the PageRank algorithm to calculate the authority of each paper.Present single or multiple papers' authority value ranking changes and citation number ranking changes.Extract basic information of each scholar,includes the number of scholars,scholar's name,scholar's productivity,index number of each scholar's works.According to the message of literature and scholars,calculate each scholar's H-index and authority.Present single or multiple scholars' H-index value ranking changes and authority value ranking changes.Finally,based on different data standards,draw six scatter plots to show scholars' distribution(3)The coauthor network module's design and actualizationConstruct the coauthor network according to coauthor relations.Calculate the density and global clustering coefficient of the coauthor network.Calculate and present each scholar's centrality,present degree centrality value ranking changes.Finally,add the citation relations to coauthor network to form a new network.On the basis of the new network,we design and actualize a way of community partition.