Clustering Scientific Publication Trends in Cultural Context Using Epidemiological Model Parameters

Epidemiology is a well-studied field in medicine and biology. The research of parameter boundaries of deterministic epidemiolog- ical models and their causal meaning is well documented. The adaptation to modeling information diffusion is a recent advance and has been shown in several studies. Previous research by the authors include modeling knowledge propagation in scientific publication using these methods. This study explores a cultural classification of scientific publication trends in 32 countries and 5 keywords from Soft Computing. Clusters are formed using parameters from a revised epidemiological SEIRE model (based on the basic SEIR—Susceptible, Exposed, Infected, Removed—model) for knowledge propagation. The cultural dynamics of the epidemiological model parameters for each country and their respective five keywords were transformed using Principal Compo- nent Analysis. The clusters were assessed by the correlation of their centroids with the epidemiological model parameters. The proposed method combines the epidemiological character of knowledge propagation with quantitative measures to classify trends.