Optimized Deep Learning with Binary PSO for Intrusion Detection on CSE-CIC-IDS2018 Dataset

Anomaly detection is a term refer to any abnormal behaviors, comprise security breaches of network. Deep Learning (DL)has proven its outperformance compared to machine learning algorithms in solving the complex problems of real-world like intrusion detection. Though, this approach need more computational resources and consumes long time. Feature selection is play significant role of choosing the best features that describes the target concept optimally during a classification process. However, when handle large number of features the selecting of such relevant features becomes a difficult task. Thus, this paper proposes using Binary Particle Swarm Optimization (BPSO) to solve the feature selection problem. Then, features selected from BPSO are evaluated on Deep Neural Networks (DNN) classifiers and the CSE-CIC-IDS2018 dataset. The result of the proposed model has shown comparable performance based on processing time, detection rate and false alarm rate comparing with other benchmark classifiers. Experimental results have shown a high accuracy of 95%.