DOI LINK: https://doi.org/10.59671/KDCMO

Paper ID:KDCMO

Volume:32

Issue:12

Title:A Computational Neural Network Framework for Fractional-Order Modeling of Buruli Ulcer and Cholera: Insights into Microbial Genetics and Pathogenesis

Abstract:This study develops a computational framework using a Levenberg-Marquardt backpropagation neural network (LMBNN) to solve the fractional-order Buruli ulcer (BU) and cholera model, integrating key aspects of microbial genetics and disease dynamics. The fractional-order model accounts for ten distinct categories, reflecting the nonlinear dynamics of infection, pathogen evolution, and host-pathogen interactions. Numerical solutions are derived using the stochastic LMBNN approach, supported by a dataset optimized with the Adam scheme. The data is partitioned into 76% for training and 12% each for validation and testing to minimize the Mean Square Error (MSE). The neural network employs a single layer with twelve neurons and a log-sigmoid activation function, offering precise approximations of disease dynamics. Validation against reference solutions shows negligible absolute error, confirming the solver's reliability. Statistical analyses further validate the model's robustness, making it a valuable tool for studying genetic resistance, microbial evolution, and the dynamics of infectious diseases.

Keywords:Cholera dynamics; Buruli ulcer; Neural networks; Fractional models; Levenberg-Marquardt; Epidemiology; Microbial genetics; Host-pathogen inter

Authors:Ghadeer Bukhari, Afaf S. Alwabli, S.I. Ali, Mohammed A. Balubaid, Jabr Aljedani, Ahmed Ramady, S. R. Mahmoud*

Paper PDF Link: View full PDF