A Metaheuristic Approach to Network Intrusion Detection
The prevention of intrusion in networks is vital; hence, an intrusion detection system is extremely desirable through potent intrusion detection mechanism. Several studies have been conducted in the domain of intrusion detections. However, some of them suffer from high false alarms, in terms of the usage of a raw dataset that contains redundancy. This paper, therefore, proposes a multi-level dimensionality reduction framework that is based on meta-heuristic optimization and principal component analysis (PCA). To achieve the aim of this research, PCA was applied for feature extraction. Genetic Algorithm and Particle Swarm Optimization, that is GA-PSO, algorithms were employed for feature selection to extract the most discriminative features to develop intrusion detection model. In the classification stage, Artificial Neural Network (ANN) and Support Vector Machine (SVM) algorithms were used to develop intrusion detection, using kddcup.data_10_percent dataset. Experimental results show that the proposed framework achieved an accuracy of 99.7% and ROC of 99.9%, while the time taken to build model is 0.23 seconds. To a very high extent, incidences of high false alarm are allayed through the GA-PSO induced feature selection method.