Advanced ANN-LMB modeling of hepatitis B transmission across sexual networks and its disability burden

Abstract

Abstract Hepatitis B virus (HBV) remains a major global health concern, with sexual transmission being a key driver among adults. This study develops a gender-stratified compartmental model of HBV spread that integrates long-term disability through gender-specific parameters. A key contribution is the integration of mechanistic modeling with artificial neural networks (ANNs), enabling efficient emulation of the model’s nonlinear dynamics. Numerical solutions from the classical Runge-Kutta 4th-order (RK4) method were used as training data for an ANN optimized with the Levenberg–Marquardt algorithm (ANN–LMB). The trained ANN accurately reproduces compartmental dynamics with minimal error and provides a fast surrogate for sensitivity exploration. Analysis of the basic reproduction number ( $$R_0$$ ) reveals the strong influence of same-sex transmission rates, contact patterns, and disability onset parameters. These findings highlight the importance of behavioral interventions, vaccination coverage, and early detection of chronic carriers. Overall, the proposed ANN–LMB framework enhances computational efficiency and offers a biologically informed approach for exploring complex HBV transmission dynamics.