A Novel DNN-Based IDS System Combined with an LR-GA Method to Detect Attacks

With the rapid expansion of the Internet of Things devices and applications, current security issues are more challenged because cyberattacks are getting more complicated and harder to spot. Hence, many intelligent attack detection approaches have been developed to face these problems in recent years. Intrusion detection systems (IDSs) based on machine learning approaches have been rising strongly due to their efficiency and scalability. However, the accuracy of attack detection by these AI-based systems is dependent on several aspects, including the features of the training dataset and the training model. This study proposes a novel intrusion detection system (IDS) based on a deep neuron network to enhance the accuracy of attack detections, especially the logistic regression (LR) optimized by the genetic algorithms (GA) method used in it to select essential features of a dataset. Our proposed model is evaluated on KDD 99 and IoT-23 datasets, bringing a better attack detection rate than other previous proposals.