Logo image
Topic-based Pagerank: toward a topic-level scientific evaluation
Journal article   Peer reviewed

Topic-based Pagerank: toward a topic-level scientific evaluation

Erjia Yan
Scientometrics, v 100(2), pp 407-437
Aug 2014

Abstract

Scientific evaluation Impact Interdisciplinary Studies Topic models Computer Science Information Storage and Retrieval PageRank Library Science
Within the same research field, different subfields and topics may exhibit varied citation behaviors and scholarly communication patterns. For a more effect scientific evaluation at the topic level, this study proposes a topic-based PageRank approach. This approach aims to evaluate the scientific impact of research entities (e.g., papers, authors, journals, and institutions) at the topic-level. The proposed topic-based PageRank, when applied to a data set on library and information science publications, has effectively detected a variety of research topics and identified authors, papers, and journals of the highest impact from each topic. Evaluation results show that compared with the standard PageRank and a topic modeling technique, the proposed topic-based PageRank has the best performance on relevance and impact. Different perspectives of organizing scientific literature are also discussed and this study recommends the mode of organization that integrates stable research domains and dynamic topics.

Metrics

11 Record Views
28 citations in Scopus

Details

InCites Highlights

Data related to this publication, from InCites Benchmarking & Analytics tool:

Web of Science research areas
Computer Science, Interdisciplinary Applications
Information Science & Library Science
Logo image