Vulnerability Analysis and Network Intrusion Detection of Application Layer Using KDD Data Set

With the advancement in technologies and use in the field of computer network, computer vulnerabilities are also been reported by most commercial scanners, and hence creating challenges to the security analyst to work upon them. To overcome or to solve such vulnerabilities many approaches had been applied. Now analysts have been trying to merge machine learning techniques to classify such vulnerabilities in real time. In this paper we have proposed an SVM-CART based technique to solve or to make system capable of classifying vulnerabilities. This technique is the hybridization of the SVM (Support Vector Machine) and CART (Classification and Regression Tree) by combining their pros and overcome the cons of other with the pros of other. The implementation of proposed techniques has been tested on KDDCUP’99 dataset.