Notice that, in our approach, when similarity is evaluated between any two articles published in two different years, all the articles published in the time interval between these two years can only be cited by the more recent article.
The resulting symmetric value of similarity between any two articles is rooted in the invariance of the hypergeometric distribution in Eq. (2) under permutation of the pair i and j, i.e., of the two quantities (d_{i}), (d_{j}).
Then, we define the probability (P_{ito j}(p^)) that there exists a citation between any two articles whose similarity is validated at the threshold (p^) as: P_{ito j}bigl(p^bigr) = frac{K(p^{M(p^. (4).
Moreover, since the analysis was based on the APS data set, the evaluation of the similarity between any two articles was restricted to the overlap between the citations the two articles made only to other articles published in the APS journals.
Specifically, we investigate empirically a large data set of articles published in the journals of the American Physical Society (APS) [16], and we measure the similarity between any two articles by drawing on, and extending, a method originally proposed by Tumminello et al. in Ref. [17, 18] that enables us to statistically validate the overlap between the bibliographies of the two articles.
The number of direct citations connecting the included studies ranged from 2 to 99 (median 15; Additional file 1: Figure S1) with a median of 2 citations between any two articles (data not shown).
(As evidenced by the suggestion that the connection between these two articles could be 'grist' for comedy. Oh my god, Dean, I can't even believe how unfunny that is).
In other words, the probability that an article i cites another article j is an increasing function of the similarity between the two articles.
Results suggest that the number citations made by one article to another is indeed an increasing function of the similarity between the two articles.
Results suggest that the probability of a citation made by one article to another is indeed an increasing function of the similarity between the two articles.
We shall therefore introduce a method for assessing the statistical significance of the overlap between the lists of references of two articles, and we shall then use the statistically validated overlap as a measure of the similarity between the two articles.
