Research On Author Influence Evaluation Index Based On Improved PageRank Algorithm

The author is not only the main body of academic exchange but also the main component of the innovation forces,playing a crucial part in promoting cultural heritage and technological innovation.With the increasing depth and breadth of researchers,as well as increasing number of researchers,the interaction between authors has become closer.The author's citation network is a complex network that embodies the mutual citation relationship between authors.The knowledge link between the network is of great significance to the dissemination of information resources and the development of cooperation among different disciplines.Based on the author's reference network,this research studies the author' s academic influence by improving the Page Rank algorithm.By sorting out the existing methods of evaluating the author' s influence by using the Page Rank algorithm,it is found that the existing researches seldom consider the factors such as the heterogeneity of paper citation time,the difference of the author,etc.Based on previous studies,this research improves the PageRank algorithm from perspectives of the heterogeneity of paper citation time and the difference of the author' s contribution and proposes an effective method to identify the author' s influence.The main research is divided into the following parts:(1)This paper summarizes the research background and significance of the author' s influence evaluation.Starting from the two aspects of network perspective and PageRank algorithm,this paper combs the domestic and foreign research related to the author' s evaluation laying a foundation for the new method.(2)To solve the problem that the temporal heterogeneity of cited papers is seldom considered in the existing research of using the PageRank algorithm to evaluate the influence of authors,a Time-PageRank author evaluation method basing on the temporal heterogeneity of cited papers is proposed.Using the author citation network,41 authors' core journal articles published in 10 fields of Library and information from 2013 to 2017 were selected as the basic data for the empirical tests.By comparing the time-PageRank method of the PageRank algorithm,it is found that the time-PageRank method has better distinguished and adjusting functions,and has a higher correlation with existing authoritative evaluation index H index and cited frequency.(3)To solve the problem that there is no reasonable allocation of contribution difference between authors with different attribution positions in co-authored papers in the study of evaluating author's influence by using Page Rank algorithm,an author evaluation method improving of PageRank named Weight-Page Rank which based on the difference of author contribution value is proposed.Meanwhile,41 authors' core journal articles published in 10 fields of Library and information from 2013 to 2017 were selected as research objects,using the author citation network to calculate the author' s Weight-PageRank value and identify highly influential authors.Through the comparative analysis with the existing author evaluation index H index,the number of articles published,the PageRank value of cited frequency and the order of signature,it is found that the Weight-PageRank method has better recognition effect on the authors with high cited frequency,cited by highly influential authors and the first order of signature.Finally,compare this study with the improved Page Rank algorithm based on the temporal heterogeneity of cited papers and the difference of contributions to the author.It is found that the two indicators have their respective emphases in the calculation and recognition effects,which should be comprehensively analyzed in practical application.