Sunday, October 28, 2012

What is h*-index?

With citation* profiles calculated for the articles of an author, we can define his h*-index. The h*-index is defined simply in the same as the h-index, just in the calculations we use the citation* values instead of citations. To be precise, the h*-index of an author at a year is H, if H is the maximum number for which he has H number articles which have a citation* index greater or equal to H in that year. The h*-index profile of an author is the time series of his h*-index values from the year he has published his first article.

For example, if an author has five articles in 2012 with citation* indices of 3.00, 2.05, 1.05, 0.54, and 1.95 in 2012 respectively,  his h*-index(2012) will be 2.

In the next blog post, an example of how to calculate the h*-index and the h*-index profile of an author will be provided.

What is citation*?

Traditionally, citation of an article is simply the number of articles that have cited that article. We introduce citation* here which is the ratio between the citations number and the age of the article (with respect to its printed publication year). For example, an article published in 2011 with three citations in 2012 will have a citation*(2012) = 3 / (2012 - 2011) = 3. In contrast, an article published in 1990 with 45 citations will have a citation*(2012) = 45 / (2012 - 1990) = 45 / 22 = 2.05. Also, citation* of an article at its publication year is equal to total number of citations it has received in that year. In other words, the denominator is set to 1 when the calculation year is the year of publication. For example, if an article is published in 2011, and it got 2 citations in that year, its citation*(2011) = 2 / 1 = 2.

One source of ambiguity in citation* could be the definition of the publication year. Many articles are published online several months before their printed version. In order to promote online publication, the printed publication year is used in the definition of the citation*. For those articles that do not have a printed version, the online publication year plus one year is used in the calculations.

The citation* index tries to spot influential articles at their time. In other words, we think that the citation indices should be time series instead of just a number at the current year.

 I will introduce the h*-index based on the citation* in the next blog post. Again, the h*-index will be a time profile representing the influence of an author over time.