Michael S. Rosenberg’s Laboratory

Computational Evolutionary Biology & Bioinformatics

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

trend h-index

The trend h-index (Sidiropoulos et al. 2007) is essentially the opposite of the contemporary h-index. It is designed to measure how current an author's impact is by how recently they are being cited. The trend score for a publication is measured as

$$S^t_i = \gamma \sum\limits_{j=1}^{C_i}{\left(Y-Y_{j.i}+1\right)^{-\delta}}$$

where γ and δ are parameters (often set to 4 and 1, respectively, just as with the contemporary h-index) and Yj.i is the year of the jth citation for publication i. If the number of citations for publication i in year Yk is Ci.k, this can also be written as $$S^t_i = \gamma \sum\limits_{k=1}^{Y}{C_{i.k}\left(Y-Y_k+1\right)^{-\delta}}.$$The trend h-index is the largest value for which an author has ht publications with at least Stht.

$$h^t = \underset{i}{\max}\left(i \leq S^t_i\right)$$

Example

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

Citations (Ci)36421411992232111000
—1997—10001
—1998—523100
—1999—13611200
—2000—582436012000
—2001—311543412210111000
Adjusted Citations (\(S^{t}_i\))134.0099.1330.0024.3324.3318.678.008.006.804.004.004.004.000.000.000.00
Rank (i)12345678910111213141516
hC = 8

The largest rank where \(i \leq S^{t}_i\) is 8.

History

Yearht
19972
19984
19995
20007
20018
200212
200314
200417
200517
200620
200723
200824
200925
201029
201129
201232
201333
201433
201532
201629
201729

References