Intelligent Network Intrusion Detection Using DT and BN Classification Techniques

Security is becoming a critical part of organizational informationsystems. Network Intrusion Detection System (NIDS) is an importantdetection that is used as a countermeasure to preserve data integrity andsystem availability from attacks. The detection of attacks against computernetworks is becoming a harder problem to solve in the field of Networksecurity. Intrusion Detection is an essential mechanism to protectcomputer systems from many attacks. The success of an intrusiondetection system depends on the selection of the appropriate features indetecting the intrusion activity. In NIDS electing unnecessary featuresmay cause computational issues and decrease the accuracy of detection. This paper describes a technique of applying Genetic Algorithm (GA) tochoose features (attributes) of KDDCUP99 Dataset. We have chosen thestandard dataset KDDCUP from MIT, U.S.A, which is used for IDSresearch oriented projects. In this paper a brief overview of the IntrusionDetection System and genetic algorithm is presented. We used BayesNetwork (BN), and Decision Tree (DT) Tree approaches for classifyingthe network attacks for chosen attribute dataset. These models gave betterperformance compared with the all features of dataset