Application of artificial neural networks to solve differential equations in Zika virus dynamics with exponential kernel derivatives

The Zika virus is a relatively new pathogen that has quickly become a major issue in virology. Infectious mosquitoes of the Aedes genus, particularly Aedes aegypti, spread the Zika virus in tropical and subtropical areas. Most bites from Aedes mosquitoes occur during the daytime. Dengue, chikungunya, and urban yellow fever are all spread by the same insects. Mathematical modeling has helped researchers better understand the mechanisms of Zika virus invasion and evaluate prospective management techniques. In this study, we suggest a unique approach to employing an Artificial Neural Network (ANN) to solve the system of differential equations describing the susceptible people, the susceptible vectors, the infected humans, the infectious vectors, and the recovered humans. Applying a unique Caputo-Fabrizio derivative yields semi-analytical findings for the model. The system of differential equations used to create the data is solved using the Laplace Adomian Decomposition (LAD) method. This makes use of a supervised learning framework for training the ANN using LAD-generated data. By comparing the ANN's projected solution to the LAD solution and determining the average error across all system functions, we can evaluate the ANN's accuracy.