An Efficient Intrusion Detection Approach Using Enhanced Random Forest and Moth-Flame Optimization Technique

Abstract

The recent advancements in the computer networks pave a sophisticated platform to the “Black hat” attackers, which poses a major challenge to network security. Intrusion detection is a significant research problem in network security which motivates the researchers to focus on the development of a robust Intrusion Detection System (IDS). Several research works in network infrastructure reveal the significance of Intrusion Detection Systems (IDS) in protecting the IT infrastructure against the ever-advancing nature of cyber attacks. In a similar manner, machine learning has a pivotal role in enhancing the performance of intrusion methodologies. Hence, this work presents Moth-Flame Optimization Algorithm-based Random Forest (MFOA-RF) to build an efficient intrusion detection methodology for networks, and it enhances the RF technique by tuning the parameter. The experiments were implemented using NSL-KDD Cup dataset and the results were validated in terms of classification accuracy and false alarm rate.