Providing a Consistent Method to Model the Behavior and Modelling Intrusion Detection Using A Hybrid Particle Swarm Optimization-Logistic Regression Algorithm

Abstract

An intrusion detection system is a collection of instruments, methods, and documentation that aid in identifying, determining, and reporting unwanted or illegal network activity. Intrusion detection systems are built as software and hardware systems, each with its own set of benefits and drawbacks. Because of the intrusion detection system's nonlinearity and nonstationary, the correctness of traditional methods, such as regression analysis and neural networks, was limited. In this research, a regression-based prediction model is proposed to handle an intrusion detection behavior problem. To develop an effective regression model, the parameters must be carefully adjusted. This present research introduces a hybrid methodology called real-value particle swarm optimization (RPSO) algorithm regression, which uses real-value particle swarm optimization algorithms to find the optimal parameters. Then, it uses the best parameters to build the regression models. The method is used to forecast the data related to an intrusion detection behavior from the VirusTotal dataset. Due to the root mean square error (RMSE) 0.0234 and the mean absolute percentage error (MAE) 1.845, the experimental results show that RPSO performs best the standard regression and backpropagation (BP) neural network models (MAPE). It was proved that the RPSO model is a practical method to recognize the behavior of the intrusion detection system feature.