PREDICTING THE TRANSMISSION DYNAMICS OF TUBERCULOSIS VIA CAPUTO FRACTIONAL ORDER MODEL WITH NEURAL NETWOR
Abstract
This study presents the development of a hybrid Fractional OrderDierential Equation (FODE) and Articial Neural Network (ANN) model designed to predict the dynamics of Tuberculosis (TB) in Nigeria. The analysisutilized data sourced from the World Health Organization (WHO) TB Database and the Nigeria category, spanning the years 2010 to 2020. The Caputoderivative was used to formulate the fractional tuberculosis model which wasenhanced with an ANN framework. The derived FODEs were discretized usingthe Gr¨unwald-Letnikov method for parameter estimation and numerical simulation of the TB data in MATLAB, employing varying memory values for thefractional-order model parameter 0 < α ≤ 1. To enhance predictive accuracy,we integrate an Articial Neural Network (ANN) with the FDE model, leveraging machine learning techniques for parameter estimation and forecasting.The ANN is trained using real-world TB data, employing the sigmoid functionto represent time-dependent transmission rates. Our results demonstrate thatthe fractional-order model provides a more exible and accurate representation of TB dynamics compared to classical integer-order models. The proposedhybrid approach eectively captures disease trends, making it a valuable toolfor epidemiological analysis and public health decision-making.
