Computational framework for hearing loss due to mumps virus model using Gudermannian neural networks and optimization techniques

This study presents a novel Gudermannian neural network (GNN) that combines global and local optimization techniques from heuristic genetic algorithms (HGAs) and sequential quadratic programming (SQP), i.e., GNNs–HGA–SQP, to address the nonlinear dynamics of hearing loss in children caused by the mumps virus (MV). This research aims to enhance our understanding of the disease’s complex behaviors through advanced computational and optimization techniques. Hearing loss is caused by genetic causes, birth problems, specific viral diseases, persistent otitis media, exposure to noise, demographic variables (study location, sex, education, race, and age), and cardiovascular disorders (hypertension, stroke, smoking status, and diabetes). The merit function is designed based on the MV model and its initial conditions. The efficiency of the proposed technique is validated by the overlap between obtained and reference solutions as well as by negligible Ab. Er values, ranging from 10 −4 to 10 −9 . Moreover, statistical evaluations based on multiple independent executions are performed to assess the efficacy of the scheme through various tests, such as Theil’s inequality coefficient (TIC), mean absolute deviation (MABD), root-mean-square error (RMSE), and semi-interquartile ranges (SIQR), which demonstrate the validity, reliability, and convergence of the integrative technique. The MV model results were obtained for values ranging from 0 to 1, with a step size of 0.1. The obtained TIC, RMSE, and MABD values ranging from 10 −8 to 10 −11 , 10 −2 to 10 −6 , and 10 −1 to 10 −6 , respectively, demonstrate the strong prediction precision, reliability, and stability of the integrative technique. Additionally, numerical simulations are provided to support our theoretical findings. The proposed neural network, which integrates global and local optimization techniques, provides an efficient and valuable framework for solving ordinary differential systems, such as the hearing-loss model due to the mumps virus, by leveraging the properties of the Gudermannian function. Future research will focus on solving more complex nonlinear and fractional-order systems and applying advanced deep learning algorithms to further improve predictive performance.