Research on Simulation of Active Network Security Structure Model Based on Machine Learning Algorithm

Active network security is a comprehensive strategy, which aims to reduce the impact of network threats on organizations by taking multi-level measures. This requires continuous efforts and investment to ensure that the network and information systems can remain secure. The traditional network security protection of security guards, antivirus software and firewalls is often single, so it is difficult to actively defend against network threats. In this paper, active defense is applied to complex networks, and a new active network security structure model is established. According to the characteristics of nonlinear time series of NSS (Network security situation) values, the model uses the advantages of SVM (support vector machine) to fit nonlinear data, and uses PSO (Particle swarm optimization) to optimize the parameters of SVM, and proposes an NSS prediction method based on PSO-SVM. The simulation results show that compared with other NSS evaluation models, PSO-SVM has the least modeling time and evaluation time, and the smallest error of the prediction model, which verifies the superiority of its prediction accuracy.