Enhanced Intrusion Detection Using RNA Encoding and Frequent Pattern Mining on CICIDS2017 Dataset

Abstract

Intrusion Detection System is a security system that serves as a shield to the infrastructure in place. Over the years, IDS technology has expanded significantly in order to meet the demands of new computer crime. Starting from the middle of the eighties to the current century, efforts have been made to improve its ability to identify attacks without compromising the performance of the network. This paper proposes a new method for a misuse IDS model based on RNA encoding and Frequent Pattern Mining for improving pattern extraction and classification, where Frequent Pattern Mining can deal with large amounts of data, therefore fast extractions of patterns and can detect intrusion in real-time. In this work, the characteristic patterns related to the malicious activity are identified using the CICIDS2017 dataset containing different types of network attacks. The proposed system comprises four stages: Data Selection and Preprocessing, RNA encoding, Frequent Pattern Mining and lastly classification. The experiments showed that the proposed IDS model has high accuracy and lower computational complexity as well as better scalability to be applied to real-time network intrusion detection.