Predictive modelling of antibacterial efficacy in TiO2/ZnO/CS nanocomposites using artificial neural networks.

The contamination of water by drug-resistant pathogens underscores the urgent need for advanced antibacterial materials. In this study, we introduce a novel ternary TiO2/ZnO/CS (CS) nanocomposite (1:2:1 ratio) that exhibits synergistic antibacterial mechanisms that combining photocatalytic reactive oxygen species (ROS) generation, Zn2+ ion release, and CS-mediated membrane disruption. Structural and surface characterization confirmed a hierarchical mesoporous scaffold, while disc diffusion tests demonstrated significantly enhanced antibacterial efficacy (inhibition zone: 6.1 ± 0.17 mm), surpassing the performance of individual components. Beyond material innovation, this work pioneers an integrated experimental-computational approach by employing an artificial neural network (ANN) as both a predictive and diagnostic tool. The ANN achieved high prediction accuracy (R2 = 0.96, MSE = 0.12) and, through sensitivity analysis, identified specific surface area and ZnO content as key determinants of antimicrobial performance. Unlike conventional statistical methods or prior ANN applications limited to photocatalysis or adsorption, our model delivers mechanistic, data-driven insights into structure to property activity relationships. This dual advancement is a highly effective antibacterial nanocomposite, and an AI-driven design methodology establishes a transformative framework for rational development and accelerated optimization of multifunctional nanomaterials for water purification.