The Expansion of the Rubricator of the VINITI RAS Abstract Journal and the List of Keywords for Epidemic Models: Bibliometric and Semantic Analysis Based on the Lens.org Database and the DeepSeek-R1 Large Language Model

The results of a quantitative bibliometric analysis of publications (2003–2023) on the mathematical modeling of epidemics, encompassing both biological infections and the spread of information in social networks, are presented. Based on the Lens.org database, a significant growth in the number of publications, peaking in 2021 due to the COVID-19 pandemic, has been revealed, with the United States and China being the leading countries. A key result is the development and validation of an expanded set of 12 key terms for the precise indexing of publications. The relevance of this set has been assessed using the DeepSeek-R1 large language model, which analyzed the contextual proximity of the selected terms to the thematic categories of leading specialized journals (based on the data from the SCImago Journal & Country Rank analytical resource) and their aims and scope descriptions. The application of the optimized term set significantly improved key metrics: Recall increased by 14.5% (to 0.87), the F1-score increased by 10.3% (to 0.86), and cosine similarity increased by 15.3% (to 0.83). Based on the conducted analysis, it is proposed to introduce new rubrics into the VINITI RAS Abstract Journal’s rubricator: Infection Spread Models and Biological Systems Modeling (within the “Biological Problems” section), as well as Epidemics in Social Networks and Social Systems Modeling (within the “Sociophysics” section).