Research On Ranking Algorithm Of Author's Academic Influence Based On Super Network

The evaluation of academic influence in large academic data is a hot research direction for researchers.It is mainly based on the relevant information published by the author to measure its influence in the academic field,such as the cooperation information of the author,the quantity of the author,the value of the literature and the influence of the periodicals,etc.It has strong application value for academic activities research.On the one hand,the effective evaluation of the academic influence of the author can make the scientific research institutions and universities seek strong partners,and provide some reference for their title evaluation and scientific research awards.On the other hand,it can provide the current research direction for young workers in a new field.The trend of future research will be guided,in addition,and the publishers can identify and cooperate with the influential writers or young researchers with potential,which will help to improve the influence of the institutions in the academic community.At present,the evaluation of academic influence is mainly divided into the statistical evaluation method and the evaluation based on network structure.Based on the statistical evaluation method,we often consider the citation frequency,the number of published documents,and the number of authors' cooperation,such as h-index.But the h-index is not sensitive to some papers cited by scientists in a large number.In addition,because the index is only for the evaluation of the current scientific research ability of researchers,the evaluation results are only undiminished,so it is unfair to young researchers.Aiming at the defects of h-index,many scholars put forward their deformation,such as g-index,R-index and so on.The evaluation methods based on network structure are mainly divided into two aspects: single layer and multilayer.Because the research of single layer network is often one-sided in the face of current huge academic data,the rise of multilayer network just makes up the shortage of single layer network.The assessment of the influence of the person opens the way.Therefore,combining the current research status and the research problems encountered,this paper proposes the author's academic evaluation model based on three layers of academic network including author citing network(ACL),author cooperative network(AAL)and paper citing network(PCL).And the paper proposes a comprehensive evaluation index M3-index,and uses an improved Leader Rank to evaluate the academic influence of the author.Secondly,this paper also analyzes the degree definition,degree centrality and characteristic vector centrality of the author's super network evaluation model,which is less involved in the current research literature.In addition,in order to verify the rationality of the super network model,this paper uses the Aminer data set to compare the sorting results of different algorithms,and proves the advantage of the comprehensive index M3-index in evaluating the author's influence.It is not only better in the ranking performance.But also it reflects the author's comprehensive scientific research ability and young scientists' potential for future development.