Hybrid Firefly and Particle Swarm Optimization Designed for Xgboosttuning Problem: Intrusion Detection in Network

Abstract

The growing of threads and intrusions on networks make the need for developing efficient and effective intrusion detection systems a necessity. Powerful solutions of intrusion detection systems should be capable of dealing with central network issues such as huge data, high-speed traffic, and wide variety in threat types. This paper proposes a feature selection method that is based on firefly algorithm, particle swarm optimization and xgboost. The proposed method improves the performance of intrusion detection by removing the irrelevant features and reduces the time of classification by reducing the dimension of data. The XGBoost model was employed to evaluate each of the feature subsets produced from firefly and particle swarm optimization technique. The main merit of the proposed method is its ability in modifying the firefly algorithm and particle swarm optimization to become suitable for selection of features. To validate the proposed approach, the popular NSL-KDD dataset was used in addition to the common measures of intrusion detection systems such as overall accuracy, detection rate, and false alarm rate. The proposed method achieved an overall accuracy of 78.89