期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2012
卷号：44
期号：2
页码：308-315
出版社：Journal of Theoretical and Applied
摘要：Literature is usually evaluated based on impact factor, cited information or the quality of literature database. Here, some more exact evaluation schemes and algorithms are investigated and a synthetically reputation model and its evaluation algorithm are proposed. The reputation model integrates much more information of academic interactions such as reviews and comments. The literature reputation can be modeled by complex relationships among literature, authors, publish sources and readers. The model is in some ways similar to the Google PageRank model. However, in the literature reputation model, the relationships belong to different classes and have different importance that is great different from PageRank which regard all links as the same class. We propose a centralized algorithm that can compute the reputation most exactly and a distributed version that can run in P2P systems. The simulations show that the reputation evaluation algorithm is convergent and efficient. The communication cost of the distributed algorithm is low if the convergence condition is relaxed.
关键词：Reputation Evaluating; Literature Evaluation; Pang Rank; Impact Factor
