A study on author ranking

Abstract

In this research we have studied the problem of ranking papers and authors based upon their citations by other authors. Simple count of number of authors citing a particular paper or author may not be very representative of the contributions made simply because a paper or authors work may be of too prohibitive complexity for too many authors to cite. It may so happen that only accomplished researchers of grater heights can understand and assimilate their deep insightful contributions. So simple count of number of citations may well fail to capture the essence. So we proposed algorithms avoiding damping factors and assigning the weight (1 􀀀 d) as has been assigned in page ranking algorithm irrespective of any citation. We have experimented with two versions of the algorithm. In the first version the score of a paper/author has been completed based on the scores of papers that have cited it. To do some justice to the papers that fail to attract citation from too many authors due to their complexity we have considered the average quality of citing papers. However, to give some weights to number of papers citing it we have multiplied the score with the square root of number of papers citing it. Experimental results of all these versions including paper rank have been presented.The current algorithms fail to properly rank authors whose citation counts are less but whose contributions are deemed important by the leading experts in the field. So in scoring of a researcher, we would like to factor in the average score of researchers who cited their work. To score these authors more fairly, we want to introduce a new Linear Programming formulation based scoring system for researchers. The purpose of Author Rank is to recognize the expertise of a person within certain subjects and what others think about the content they publish. We want to use an algorithm which is based on the System of linear equations. And the closest algorithm of this is Google’s Page Rank algorithm. There is a need to consider the credibility of each author in order to examine the relativity of this broad data.