Performance Analysis of Data Mining Approaches in Intrusion Detection
Authors:
- P. AmudhaPAP. Amudha
- H. Abdul RaufHRH. Abdul Rauf
- 1 July 2011
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Abstract
Intruder is one of the most publicized threats to security. In recent years, intrusion detection has emerged as an important technique for network security. Data mining techniques have been applied as a new approach for intrusion detection. The quality of the feature selection methods is one of the important factors that affect the effectiveness of Intrusion Detection system (IDS). This paper evaluates the performance of data mining classification algorithms namely J48, Naive Bayes, NBTree and Random Forest using KDD CUP'99 dataset and focuses on Correlation Feature Selection (CFS) measure. The results show that NBTree and Random Forest outperforms other two algorithms in terms of predictive accuracy and detection rate.Keywords
This publication has 8 references indexed in Scilit:
- Intrusion detection using data mining techniques2010
- Improving Effectiveness of Intrusion Detection by Correlation Feature Selection2010
- Machine Learning Techniques for Feature Reduction in Intrusion Detection Systems: A Comparison2009
- Data mining-based intrusion detectorsExpert Systems with Applications, 2008
- Attributes Reduction Based on GA-CFS Method2007
- Random ForestsMachine Learning, 2001
- Wrappers for feature subset selectionArtificial Intelligence, 1997
- An Intrusion-Detection ModelIEEE Transactions on Software Engineering, 1987