| Finding experts with influential papers or a list of outstanding scholars, is particularly important to understand the development of the related areas. Besides, measuring the contribution of each scholar can give respect to the influential experts as well as provide a standard for academic awards. In this context, many ranking methods are proposed to measure the academic impact of scholars. Citation based indexes and homogeneous based models are the traditional methods. However, researchers have proven that the employment of hybrid or heterogeneous networks has a better performance with capturing the complex research communications and interactions.Heterogeneous networks, which contain richer academic information, are more and more popular in the researches of author impact. These methods calculate the author impact by the entities linked to them through a random walk process. However, when calculating author impact in heterogeneous networks, current researches take all the nodes as equal at the beginning of the process. In reality, the papers are of different importance and their starting values should be distinguished as a matter of course. In this paper, we firstly propose to value the papers with their PageRank scores, which are obtained in paper citation network. We propose the TAPRank model, which measures author impact in heterogeneous network with papers valued by their PageRank scores. In addition, we think authors with recent publications and citations are more active and should get more academic impact in the research area. We use a time-aware PageRank algorithm to give more value to newly published papers and recently happened citations.To prove the applicability of our proposed model, we take the experiment on two datasets which are the DBLP dataset and the APS dataset. They are quite different from each other because the characters of each domain. The results of our model show a better performance than other state-of-the-art methods. |
PDF Full Text Request