AI-DRIVEN PARAMETER LEARNING FOR HIGH-DIMENSIONAL FRACTIONAL SEIR MODELS

Abstract

This study presents a fractional-order SEIR epidemic model consisting of nine compart- ments for the analysis of complex disease transmission dynamics. The proposed model is formulated using the Atangana–Baleanu Caputo fractional derivative in order to incorporate memory and heredi- tary effects that arise in epidemic processes. A data-assisted computational methodology is employed to estimate both the epidemiological parameters and the fractional order of the system. To efficiently solve the associated inverse problem, a metaheuristic optimization strategy is adopted. Numerical simulations carried out in a Python environment demonstrate that fractional memory effects contribute to reducing infection peaks and significantly influence the temporal evolution of disease spread. The proposed framework provides valuable insight for epidemic forecasting and computational disease modeling.