Modeling Virus Spread in Computer Networks: An Extended SEIR Approach Using Artificial Neural Networks with Levenberg-Marquardt Algorithm

This paper introduces a prolonged SEIR epidemic model in computer networks that addresses the transmission dynamics of a virus and how it can be controlled. By employing artificial neural networks (ANNs) and training them using the Levenberg-Marquardt (LM) algorithm, we study the behavior of the model and forecast virus spreading in the network. We find the basic reproduction number (R0) of the model and provide essential insight into the threshold for virus persistence or eradication. By stability analysis, we verify that the model has equilibrium at both disease-free and endemic states, which is crucial in designing efficient antivirus measures. The research tests the performance of the model with three split data cases, namely 60-20-20, 70-15-15, and 80-10-10, to assess the training accuracy vs. generalization trade-off. Numerical methods used through MATLAB enable us to compute graphical solutions illustrating the long-term behavior of virus diffusion. This study deepens our knowledge of virus transmission in computer networks and facilitates the creation of stronger defenses against network-based cyber-attacks.