Anomaly detection with multinomial logistic regression and Naïve bayesian

Abstract

Intrusion Detection by automated means is gaining widespread interest due to the serious impact of Intrusions on computer system or network. Several techniques have been introduced in an effort to minimize up to some extent the risk associated with Intrusion attack. In this paper, we have used two novel Machine Learning techniques including Multinomial Logistic Regression and Naïve Bayesian in building Anomaly-based Intrusion Detection System (IDS). Also, we create our own dataset based on four attack scenarios including TCP flood, ICMP flood, UDP flood and Scan port. Then, we will test the system's ability of detecting anomaly-based intrusion activities using these two methods. Furthermore we will make the comparison of classification performance between the Multinomial Logistic Regression and Naïve Bayesian. © 2013 Springer Science+Business Media Dordrecht(Outside the USA).