Increasing Accuracy of Intrusion Detection System using Support Vector Machine and Bat Algorithm

Intrusion Detection System (IDS) is a device or software to monitor networks or systems in order to search for malicious activity or policy violations. The most critical challenge in IDS is distinguishing between normal and malicious traffic. The accuracy of IDS has been recently improved through various Learning models. However, the accuracy of the IDS systems still remained a challenge, as attacker and malicious nodes change their behaviors frequently. This research proposes a model to increase the accuracy of the IDS using Support Vector Machine (SVM) and Bat Algorithm (BA). SVM is one of the machine learning algorithms that recently applied by researches for various classification and regression problems and it shows an outstanding performance. Anyhow, the performance of SVM strongly depends on its parameters. In this research, we use BA to optimize the parameters of SVM to increase the accuracy of the IDS. BA has a distinct advantage over other metaheuristic algorithms, BA has the capability of automatically zooming into a region where promising solutions have been found. We evaluate the propose model using Nsl-Kdd dataset. The comparison between one of the recent machine learning algorithms and the present study indicates that the propose model has higher accuracy and better performance than the previous models.