Modified Naive Bayes Algorithm for Network Intrusion Detection based on Artificial Bee Colony Algorithm

Abstract

Intrusion detection algorithm based on machine learning is a research hotspot in network security detection. The diversity of network intrusion detection data sets is one of the major factors that affect the practical application of machine learning. Therefore, some swarm intelligence algorithms were utilized to optimize parameters of machine learning methods for feature selection or feature weight in network intrusion. In this paper, a modified Naive Bayes algorithm based on artificial bee colony algorithm (ABCWNB, in brief) is proposed. The proposed method is tested on two public data sets and NSL-KDD data sets. Experimental results show that compared to Naive Bayes classifier based on genetic algorithm (GAWNB), Naive Bayes classifier based on grey wolf optimizer (GWOWNB), Naive Bayes classifier based on water wave optimization (WWOWNB) and basic Naive Bayes classifier, the proposed method can effectively improve the network intrusion detection rate, which can well detect various types of network intrusion and greatly improve the security performance of the network.