A two-hidden-layer Levenberg–Marquardt neural network for supervised trajectory approximation of a nonlinear anthrax in animals epidemic model

The purpose of this investigation is to present the computational capabilities for the anthrax infection in animals (AIA) that divide the populations into infected, susceptible, vaccinated, and recovered groups using a two-layered neural network approach. The construction of a neural network is obtained by using double hidden layers using the sigmoid function in both layers together with 15 and 25 neurons. A two-layer neural network outclasses a single-layer network to capture the intricate and nonlinear inputs and outputs relations, which permit better demonstrating of complicated designs. In comparison with the one-layered networks, double layer frameworks can acquire classified data representations, extracting fundamental topographies in layer 1 and linking them to design higher-level topographies in layer 2 that are used to develop performance for intricate systems. A Levenberg Marquardt backpropagation is used in optimization that is appropriate for nonlinear systems due to its effective optimization, which regulates the step size and direction to reduce the cost function. A numerical Euler method is applied to generate the dataset, which minimizes the MSE by separating the statistics into justification as 15%, testing as 16%, and training as 69%. The similarity of the findings and a minor absolute error are used to determine the accuracy of the suggested algorithm, while the reliability of the solver is observed through different tests.