Network Intrusion Detection Using Machine Learning for Virtualized Data

Abstract

AbstractNetwork uses the intrusion detection system when it has to examine the malicious activity in it. In this process, the system is scanned by the intrusion detection system which is a device or software for detecting or identifying the malicious activity. Security mechanisms are required for the substantial growth in the number of applications in order to maintain their protection. There are different types of intrusion detection methods like anomaly based and signature based, but most trending subject to the researchers is machine learning (ML)-based methods. Accuracy is the most influenced parameter for intrusion detection performance. False alarm reduction and detection rate increment or detecting time decrement can be achieved with improvement in the intrusion detection system accuracy. A network intrusion detection using ML for virtualized data is proposed in this paper. Better accuracy is simultaneously increasing the performance. Intrusion detection system main work is huge network traffic data analysis. The classification problems can be solved by machine learning techniques such as Naïve Bayes, random tree, and support vector machine (SVM). NSL-KDD (knowledge discovery dataset) is used for evaluation of proposed intrusion detection system. To perform comparative analysis, average detection time, misclassification rate, and accuracy are calculated.KeywordsMalwareNSL-KDDMachine learningIntrusion detectionNaïve BayesSupport vector machine