1. Rooted citation graphs density metrics for research papers influence evaluation.
- Author
-
Giatsidis, Christos, Nikolentzos, Giannis, Zhang, Chenhui, Tang, Jie, and Vazirgiannis, Michalis
- Subjects
CITATION analysis ,GRAPH theory ,BIG data ,INFORMATION theory ,MEASURE theory - Abstract
• Novel impact metrics based on graph density. • Impact evaluation on large data sample. • Novel results not correlated with established metrics • Validation with commonly accepted ground-truth information. Traditionally, citation count has served as the main evaluation measure for a paper's importance and influence. In turn, many evaluations of authors, institutions and journals are based on aggregations upon papers (e.g. h-index). In this work, we explore measures defined on the citation graph that offer a more intuitive insight into the impact of a paper than the superficial count of citations. Our main argument is focused on the identification of influence as an expression of the citation density in the subgraph of citations built for each paper. We propose two measures that capitalize on the notion of density providing researchers alternative evaluations of their work. While the general idea of impact for a paper can be viewed as how many researchers have shown interest to a piece of work, the proposed measures are based on the hypothesis that a piece of work may have influenced some papers even if they do not contain references to that piece of work. The proposed measures are also extended to researchers and journals. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF