A MACHINE LEARNING APPROACH WITH VERIFICATION OF PREDICTIONS AND ASSISTED SUPERVISION FOR A RULE-BASED NETWORK INTRUSION DETECTION SYSTEM

José Ignacio Fernández-Villamor, Mercedes Garijo

2008

Abstract

Network security is a branch of network management in which network intrusion detection systems provide attack detection features by monitorization of traffic data. Rule-based misuse detection systems use a set of rules or signatures to detect attacks that exploit a particular vulnerability. These rules have to be hand-coded by experts to properly identify vulnerabilities, which results in misuse detection systems having limited extensibility. This paper proposes a machine learning layer on top of a rule-based misuse detection system that provides automatic generation of detection rules, prediction verification and assisted classification of new data. Our system offers an overall good performance, while adding an heuristic and adaptive approach to existing rule-based misuse detection systems.

Paper Citation

in Harvard Style

Ignacio Fernández-Villamor J. and Garijo M. (2008). A MACHINE LEARNING APPROACH WITH VERIFICATION OF PREDICTIONS AND ASSISTED SUPERVISION FOR A RULE-BASED NETWORK INTRUSION DETECTION SYSTEM . In Proceedings of the Fourth International Conference on Web Information Systems and Technologies - Volume 1: WEBIST, ISBN 978-989-8111-26-5, pages 143-148. DOI: 10.5220/0001524801430148

in Bibtex Style

@conference{webist08,
author={José Ignacio Fernández-Villamor and Mercedes Garijo},
title={A MACHINE LEARNING APPROACH WITH VERIFICATION OF PREDICTIONS AND ASSISTED SUPERVISION FOR A RULE-BASED NETWORK INTRUSION DETECTION SYSTEM},
booktitle={Proceedings of the Fourth International Conference on Web Information Systems and Technologies - Volume 1: WEBIST,},
year={2008},
pages={143-148},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001524801430148},
isbn={978-989-8111-26-5},
}

in EndNote Style

TY - CONF
JO - Proceedings of the Fourth International Conference on Web Information Systems and Technologies - Volume 1: WEBIST,
TI - A MACHINE LEARNING APPROACH WITH VERIFICATION OF PREDICTIONS AND ASSISTED SUPERVISION FOR A RULE-BASED NETWORK INTRUSION DETECTION SYSTEM
SN - 978-989-8111-26-5
AU - Ignacio Fernández-Villamor J.
AU - Garijo M.
PY - 2008
SP - 143
EP - 148
DO - 10.5220/0001524801430148