Your browser version is outdated. We recommend that you update your browser to the latest version.

Manuscript Title: Intelligent Network Intrusion Detection Using DT and BN Classification Techniques

Author : P. Srinivasulu, R. Satya Prasad and I. Ramesh Babu

Email : srinivasulupamidi@yahoo.co.inprofrsp@gamil.com, csnu@nagarjunauniversity.ac.in

Abstract: Security is becoming a critical part of organizational information systems. Network Intrusion Detection System (NIDS) is an important detection that is used as a countermeasure to preserve data integrity and system availability from attacks. The detection of attacks against computer networks is becoming a harder problem to solve in the field of Network security. Intrusion Detection is an essential mechanism to protect computer systems from many attacks. The success of an intrusion detection system depends on the selection of the appropriate features in detecting the intrusion activity. In NIDS electing unnecessary features may cause computational issues and decrease the accuracy of detection. This paper describes a technique of applying Genetic Algorithm (GA) to choose features (attributes) of KDDCUP99 Dataset. We have chosen the standard dataset KDDCUP from MIT, U.S.A, which is used for IDS research oriented projects. In this paper a brief overview of the Intrusion Detection System and genetic algorithm is presented. We used Bayes Network (BN), and Decision Tree (DT) Tree approaches for classifying the network attacks for chosen attribute dataset. These models gave better performance compared with the all features of dataset.

Keywords: Genetic Algorithms, Classification methods, Confusion matrix, Fitness function, and Crossover.

Vol 2 (1)