MODELING AND FORECASTING TWO-STRAIN DISEASE AMPLIFICATION DYNAMICS USING MACHINE LEARNING APPROACHES

A two-strain disease model incorporating amplification is developed to estimate the prevalence of drug-resistant ([Formula: see text]) and drug-sensitive ([Formula: see text]) strains. Drug-susceptible strains transition to drug-resistant ones due to factors like insufficient treatment or amplification. In this context, drug-resistant and drug-susceptible strains are analyzed together. Adopting a fractal-fractional dynamics model based on the Caputo derivative, the study seeks to explore the dynamics of the two-strain disease. Theoretical conclusions are drawn using fixed-point theory, and the system’s stability is evaluated through the Ulam–Hyers method. Numerical simulations are conducted using the fractional Adams–Bashforth approach. Pre-existing data are used to simulate the system by analyzing the different values that determine both the fractal dimension and fractional order. Several graphical illustrations are provided to interpret the system’s behavior across different fractional orders and dimensions. The dataset is segmented into training, testing, and validation sets, with Artificial Neural Network (ANN) techniques applied for detailed examination. Each category undergoes a thorough analysis to provide insights into the system’s dynamics.