Predicting author h-index using characteristics of the co-author network
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
The objective of this work was to test the relationship between characteristics of an author's network of coauthors to identify which enhance the h-index. We randomly selected a sample of 238 authors from the Web of Science, calculated their h-index as well as the h-index of all co-authors from their h-index articles, and calculated an adjacency matrix where the relation between co-authors is the number of articles they published together. Our model was highly predictive of the variability in the h-index (R 2 = 0.69). Most of the variance was explained by number of co-authors. Other significant variables were those associated with highly productive co-authors. Contrary to our hypothesis, network structure as measured by components was not predictive. This analysis suggests that the highest h-index will be achieved by working with many co-authors, at least some with high h-indexes themselves. Little improvement in h-index is to be gained by structuring a co-author network to maintain separate research communities. 2012 Akadmiai Kiad, Budapest, Hungary.