An Improved Feedforward Neural Network Using Salp Swarm Optimization Technique for the Design of Intrusion Detection System for Computer Network

Due to the drastic increase in the rate of cyberattacks, network security has become the highest priority in the recent technological era. As the cyberattacks have become more sophisticated in nature, it emphasizes the need for a second line of defense called the Intrusion Detection System (IDS). In recent years, computational-intelligence-based IDS model has become a major choice of researchers for minimizing the conflicts between the high detection rate and less false alarm rate. In this way, we present a Salp Swarm Optimization based Feedforward Neural Network (SSO-FFN) to build an intrusion detection methodology for computer networks. The performance of SSO-FFN over the state-of-the-art intrusion detection methodologies was validated using NSL-KDD cup intrusion dataset with regards to detection rate and false alarm rate. This paper proposes a technique which has a better accuracy, less false alarm rate, and high detection rate when compared with the existing techniques.