Prediction of citation dynamics of individual papers

We apply stochastic model of citation dynamics of individual papers developed in our previous work (M. Golosovsky and S. Solomon, Phys. Rev. E\textbf{ 95}, 012324 (2017)) to forecast citation career of individual papers. We focus not only on the estimate of the future citations of a paper but on the probabilistic margins of such estimate as well.

💡 Research Summary

The paper presents a probabilistic framework for forecasting the citation trajectory of individual scientific papers and quantifies the inherent uncertainty of such forecasts. After reviewing the literature, the authors distinguish two major approaches to citation prediction: (1) models that rely on a priori features (title, author reputation, journal prestige, etc.) and typically employ machine‑learning techniques on large historical datasets, and (2) models that focus on a posteriori information, chiefly the early citation history of a paper, and attempt to capture the underlying citation dynamics. The authors adopt the second approach, extending the stochastic citation model they introduced in 2017.

In the model, the instantaneous citation rate λ_j(t) of paper j at time t after publication is decomposed into a direct component λ_dir and an indirect (self‑exciting) component λ_indir. Formally, \