Identifying Important Scholars Via Directed Scientific Collaboration Networks

Jianlin Zhou,An Zeng,Ying Fan,Zengru Di

DOI: https://doi.org/10.1007/s11192-017-2619-0

IF: 3.801

2017-01-01

Scientometrics

Abstract:Scientific collaboration plays an important role in the knowledge production and scientific development. Researchers have investigated numerous aspects of scientific collaboration by constructing scientific collaboration networks. And we can perform node centrality analysis on the scientific collaboration networks to identify important scholars. In these collaboration networks, two scientists are linked if they have coauthored at least one paper and the way of constructing these networks is based on the assumption that each author’s contribution to an article is the same. However, the authors’ contributions to an article are unequal in reality and we should pay attention to the impact of this unequal credit allocation on the understanding of scientific collaboration. In this paper, we regard the first author as the most important contributor to an article and build a directed scientific collaboration network. Then we identify important scholars by analyzing this directed network. For one thing, we investigate the difference between the undirected and directed scientific collaboration network in network properties and centrality analysis. For another, we apply different centrality indices: betweenness, PageRank, SIR and HITS to the directed scientific collaboration network. As a result, we find that each indicator has a different performance and the PageRank algorithm and SIR show highly positive correlation with in-degree. The HITS algorithm also shows better property which can hep us distinguish potential young scholars and identify important collaborators.