Autores

Jorge Cruz-Trani, Ismael Robles-Martínez, Daniela Aguirre-Guerrero, Ricardo Marcelín-Jiménez.

Abstract

This work proposes a methodology to characterize the propagation factor of viral topics in complex scientific collaboration networks, where nodes are authors and edges represent co-authorships. Topic diffusion is modeled using compartmental epidemiological models (SI, SIS, SIR, and extensions), estimating the basic reproduction number $R_0$. The models are formulated as systems of parametric ordinary differential equations; the transmission ($\beta$) and recovery ($\gamma$) parameters are estimated using non-linear least squares, numerically integrated with the 4th-order Runge-Kutta (RK4) method. For empirical validation, we use annual historical data of publications indexed in the Scopus platform, previously classified into topics. Finally, the proposed methodology allows quantifying the virality of scientific topics and offers a predictive tool to detect emerging trends in academic communities.

Palabras Clave

complex networks, co-authorship networks, topic propagation, epidemiological models