A time-aware citation-based model for evaluating scientific products: extended abstract

Recently a citation-based model for ranking scientific journals and papers together with authors has been proposed [4, 5]. In that model, papers, authors and journals mutually contribute to the attribution of ranking score to each other. The rank of each subject is computed as the stationary distribution of a suitable Markov chain. In this paper, we add into the model a factor accounting for the aging of papers. In particular, the importance of each paper slowly decreases as time elapses unless a fresh citation arrives conferring new importance to the cited papers. The experimental part shows the effectiveness of the introduction of the time into the model. In fact, new papers gain importance with respect to older ones, sustaining in this way new trends of research with respect to subjects popular years ago.