Modeling the Global Citation Network using the Scalable Agent-based Simulator for Citation Analysis with Recency-emphasized Sampling (SASCA-ReS)

The paper operates within the scientometrics paradigm, specifically focusing on the computational modeling of citation network dynamics. This paradigm seeks to explain the large-scale structure and evolution of scientific literature through generative models based on principles like preferential attachment, recency, and fitness. The research puzzle is to create synthetic networks whose statistical properties match those of real-world citation data, thereby allowing for hypothesis testing and the exploration of underlying citation behaviors.

This paper is classified as Model Drift because it identifies and addresses significant anomalies between existing models and empirical observations, without overthrowing the core paradigm. The authors note that "extant theories of citation do not offer much in the way of quantitative explanation" and that previous models inadequately captured the effect of "recency." Their contribution, the SASCA-ReS simulator, is a sophisticated new instrument designed to patch this deficiency. This effort to refine and re-tool the paradigm in response to its growing inability to quantitatively match real-world data represents a drift from routine normal science puzzle-solving.