Two common ways to measure the "output" of a researcher (or research group) are to count numbers of publications and to count the citations (references to these publications in publications of others). These simple methods are flawed because they cannot easily take into account the differences in publication and citation habits in different scientific communities. An alternative approach is to view citations as hypertext links, and to use or adapt hypertext metrics to compare the scientific output of researchers, in comparison to that of others in areas with similar publication and citation patterns. We show how hypertext metrics, introduced by Botafogo, Rivlin and Shneiderman, can be modified in order to identify comparable research fields based on their publication and citation pattern. An author's performance can then be compared to that of others in research fields with a similar pattern.
[1]
J. W. de Vocht.
Experiments for the characterization of Hypertext-structures
,
1994
.
[2]
Ben Shneiderman,et al.
Structural analysis of hypertexts: identifying hierarchies and useful metrics
,
1992,
TOIS.
[3]
R. Philips.
Pseudo-metrics on bibliographic entities
,
2005,
Scientometrics.
[4]
Ronald Rousseau,et al.
The Gozinto theorem: Using citations to determine influences on a scientific publication
,
1987,
Scientometrics.
[5]
Jon M. Kleinberg,et al.
Inferring Web communities from link topology
,
1998,
HYPERTEXT '98.