Discovering author impact: A PageRank perspective
Introduction
In bibliometrics, the number of citations is an indicator used to measure the impact of scientific publications. Authors whose publications have been intensively cited usually have a higher academic impact in their respective fields; however, there are situations where citations do not provide a full perspective on the impact of an author.
Coauthorship network analysis, with its sound theory and methodology derived from physics, mathematics, graph theory, and social sciences, is expected to serve as the complement to traditional citation analysis. Specifically, the micro-level metrics for coauthorship network analysis can inform us about the power, stratification, ranking, and inequality in social structures (Wasserman & Faust, 1994). Such an approach captures the features of the individual actors in a network along with consideration of the topology of the network.
Among many indicators, PageRank has great potential for coauthorship network analysis. The PageRank algorithm (Brin & Page, 1998) assumes web hyperlinks to be trust votes and ranks the search results based on these links interconnecting them. PageRank brings a new method to information retrieval for a better ranking of the web. For coauthorship networks, the PageRank algorithm gives higher weights to the authors who collaborate with different authors, and also to authors who collaborate with a few highly coauthored authors. PageRank is thus chosen as a complementary method to citation analysis, enabling us to identify author impact from a new perspective.
The field of informetrics was selected for this study, since it is a fast-developing discipline (Bar-Ilan, 2008). More importantly, it is a coherent field, in that the selection will not result in too many breakages of collaboration ties, which is of vital importance for the robustness of coauthorship networks. In the first section, related studies on the PageRank algorithm for bibliometrics are introduced, and the second section, PageRank algorithm as well as the weighted PageRank algorithm is presented. In Section 4, we: (1) calculate the correlation coefficient between PageRank ranks and citation ranks for authors in the network, (2) discuss PageRank values under different damping factors from 0.15 to 0.85 with a 0.1 increment, (3) rank authors with the weighted PageRank algorithm, and (4) compare PageRank with the h-index, citation and PC members.
Section snippets
Related studies
Coauthorship networks can illustrate authors’ social capital in terms of collaboration in the chosen discipline (Yan & Ding, 2009). For example, through centrality studies, authors with high betweenness centrality have more opportunities to broker the flow of information and have a higher social capital (Burt, 2002). Yan and Ding (2009) found that degree centrality, closeness centrality, and PageRank also measure authors’ impacts on the field as well as their social capital. Disciplinarity can
Data in the study
The term “informetrics” was introduced by Blackert, Siegel, and Nacke in the 1970s and gained popularity by the organization of the international informetrics conference in 1987 (Egghe & Rousseau, 1990). The field of informetrics, actually, started in the first half of the 20th century with works by Lotka, Bradford, and Zipf (Egghe, 2005).
Tague-Sutcliffe (1992) defines informetrics as “the study of the quantitative aspects of information in any form, not just records or bibliographies, and in
Correlation between PageRank and citation rank
Before calculating the correlation between PR and citation rank, we present the concept of component. In social network analysis, connected graphs are referred to as components. A component of a graph is a subset with the characteristic that there is a path between one node and any other nodes in the same subset (Nooy, Mrvar, & Batagelj, 2005). A coauthorship network generally consists of many disconnected components, and we usually focus on the largest components, since metrics like closeness
Conclusion
The current study applies PR to coauthorship network analysis. The PR algorithm provides a meaningful extension to the traditionally used citation counts for authors.
Through the correlation analysis between PR and citation for each author, we find that PR and citation are correlated and that PR, to a certain degree, also measures an author’s academic impact. But they also differ when comparing to the significant correlation coefficients of paper citation networks and journal citation networks.
Acknowledgements
The authors are indebted to Ronald Rousseau and Liming Liang who provided valuable conference data to this study.
References (32)
Informetrics at the beginning of the 21st century – A review
Journal of Informetrics
(2008)- et al.
The anatomy of a large-scale hypertextual web search engine
Computer Networks and ISDN Systems
(1998) - et al.
Finding scientific gems with Google’s PageRank algorithm
Journal of Informetrics
(2007) - et al.
Refining dermatology journal impact factors using PageRank
Journal of the American Academy of Dermatology
(2007) Expansion of the field of informetrics: Origins and consequences
Information Processing and Management
(2005)- et al.
Co-authorship networks in the digital library research community
Information Processing and Management
(2005) - et al.
Bringing PageRank to the citation analysis
Information Processing and Management
(2008) - et al.
Citation influence for journal aggregates of scientific publications: Theory, with application to the literature of physics
Information Processing and Management
(1976) - et al.
A generalized comparison of graph-based ranking algorithms for publications and authors
Journal of Systems and Software
(2006) An introduction to informetrics
Information Processing and Management
(1992)
Journal status
Scientometrics
The social capital of structural holes
PageRank for ranking authors in co-citation networks
Journal of the American Society for Information Science and Technology
Introduction to informetrics: Quantitative methods in library, documentation and information science
PageRank for bibliographic networks
Scientometrics
Cited by (137)
A reviewer-reputation ranking algorithm to identify high-quality papers during the review process
2024, Expert Systems with ApplicationsMinimizing the misinformation concern over social networks
2024, Information Processing and ManagementNatural zeolite as a supplementary cementitious material – A holistic review of main properties and applications
2023, Construction and Building MaterialsThe spatial difference of multi-layer climate change information flow and network construction: A comparison of “dual carbon” scenarios
2023, Journal of Cleaner ProductionThe power of big data analytics over fake news: A scientometric review of Twitter as a predictive system in healthcare
2023, Technological Forecasting and Social Change