Predicting future citation behavior

Abstract

In this article we further develop the theory for a stochastic model for the citation process in the presence of obsolescence to predict the future citation pattern of individual papers in a collection. More precisely, we investigate the conditional distribution-and its mean-of the number of citations to a paper after time t, given the number of citations it has received up to time t. In an important parametric case it is shown that the expected number of future citations is a linear function of the current number, this being interpretable as an example of a success-breeds-success phenomenon.