A Neural Network Process for Mathematical Model of Infectious Disease Transmission System

ABSTRACT The present study provides the solutions to the infectious disease dynamical system using artificial neural network process. The disease dynamics is divided into seven groups of individuals: susceptible S ( t ), exposed E ( t ), infected I ( t ), quarantined Q ( t ), recovered R ( t ), deceased D ( t ), and vaccinated V ( t ). The proposed neural network solver is structured based on 20 neurons in the hidden layer and sigmoid activation function together with the training of Bayesian regularization. A dataset is constructed using the Adams‐Bashforth method, which is used to reduce the mean square error by data separating into training (78%), testing (12%), and validation (10%). The solutions of the model are presented in three different cases, while the accuracy is observed through the matching of the solutions, negligible absolute error, and best training standards. Moreover, the reliability and consistency of the proposed solver are observed through different tests including error histogram, transition state, regression coefficient performances, and fitness function values.