Performance Analysis of Network Intrusion Detection Systems using J48 and Naive Bayes Algorithms

Abstract

Any malicious activity on the network needs to be detected immediately to protect the user data. This helps to ensure Confidentiality, Availability, and Integrity. Machine learning algorithms are efficient tools that can be used in anomaly detection techniques to detect attacks against network. Decision Trees and Naive Bayes algorithms are the two important algorithms that can detect zero-day attacks with a great precision. While both are used for same purpose, these algorithms may produce different detection performance results on same set of data. This paper evaluates the Intrusion detection performance of these two algorithms on CIDDS-02 data set using various parameters of interest.