Michael S. Rosenberg’s Laboratory

Computational Evolutionary Biology & Bioinformatics

E-mail: msr@asu.edu
← Back to introduction

hi-index

The hi-index (Batista et al. 2006) is a simple correction of the h-index for multi-authored publications. This index is simply the h-index divided by the average number of authors in the core publications, or

$$h_i=\frac{h}{\frac{\sum\limits_{i=1}^{h}{A_i}}{h}}=\frac{h^2}{\sum\limits_{i=1}^{h}{A_i}}.$$If every publication in the core is solo-authored then hi = h. This can be an extremely harsh correction. A single core publication with a large number of co-authors may skew the average and thus lower an author's impact factor tremendously. Use of the median rather than the mean might be a fairer approach.

Example

Publications are ordered by number of citations, from highest to lowest.

Citations (Ci)42361411993222111000
Rank (i)12345678910111213141516
h = 6
Authors (Ai)3332414411421421

The h-index is 6 and the sum of the authors for publications in the core is 16, thus hi = 2.2500.

History

Yearhi
19970.3333
19980.8000
19991.1250
20001.6667
20012.2500
20022.7222
20033.5217
20043.1026
20054.0238
20064.0179
20074.8983
20085.6406
20097.5571
20107.7838
20118.7792
20127.7553
20138.0865
20148.3252
20159.2480
20169.4961
20179.4961

References