Artificial neural network-driven modeling of Ebola transmission dynamics with delays and disability outcomes.

This study develops an Artificial Neural Network (ANN)-based framework to model the transmission dynamics and long-term disability outcomes of Ebola Virus Disease (EVD). Building on existing deterministic SEIR models, we extend the framework by introducing a disability compartment, capturing the progression of Ebola survivors to chronic health complications, such as post-Ebola syndrome. The proposed model stratifies the population into various epidemiological states, incorporating delays to better reflect the natural progression and intervention strategies associated with EVD. Fundamental properties of the model, such as positivity, boundedness, and stability, have been thoroughly examined. By leveraging the Levenberg-Marquardt backpropagation (LMB) algorithm, the ANN is trained on data generated through the Runge-Kutta method to solve a system of delay differential equations (DDEs) representing disease progression. This approach offers an alternative to conventional numerical solvers, addressing limitations such as computational overhead and approximation errors. The ANN model divides the dataset into 85% training, 10% validation, and 5% testing, ensuring reliable predictions with minimal absolute error. Comparative analysis against traditional methods highlights the advantages of the ANN-based solver in handling complex, delay-integrated systems. Our results underscore the utility of integrating ANN approaches in epidemic modeling, providing insights into both short- and long-term dynamics of Ebola outbreaks. By capturing disability outcomes, this work offers a robust framework for planning healthcare interventions and optimizing resource allocation for survivor rehabilitation. The findings contribute to the development of more comprehensive models for understanding and managing infectious diseases with long-term impacts.