An Effective and Intelligent Intrusion Detection System using Deep Auto-Encoders

Abstract

Cyber threats or attacks are major security concern with the monumental growth and increased number of Internet connecting devices. Furthermore, the attackers exploit sophisticated attacks and persist for long period of time easily. The dynamic nature and large volume of cyber attacks require a responsive, adaptive and scalable protective mechanism. Many supervised and unsupervised learning methods have been developed from the domain of machine learning and data mining algorithms to detect and classify those attacks. The main aim of this proposed work is to examine the suitability of deep learning approach for intrusion detection system. In this work, IDS model is proposed based on Deep Auto Encoders (DAE). This model is trained and tested with NSL_KDD data set. The performance of the proposed system is compared with conventional machine learning algorithms namely, Logistic Regression (LR), Naïve Bayes (NB), K-Nearest Neighbor (KNN), Decision Tree (DT) and Random Forest (RF) methods. The experiment results show the improved accuracy of DAE in intrusion detection in comparison with classical machine learning algorithms.