Intrusion detection by pipelined approach using conditional random fields and optimization using SVM

Abstract

The rapid increase in network traffic and attacks made the Intrusion Detection Systems to fail in terms of accuracy and efficiency in many situations. In this paper we have proposed an approach for Intrusion Detection by Pipelined approach using Conditional Random Fields and Optimization using Support Vector Machine. The main goal of this approach in Intrusion Detection System is to achieve high accuracy and efficiency. The accuracy is maintained through the Pipelined approach and Conditional Random Fields and the efficiency is achieved through SVM. The proposed Intrusion Detection System can be used to build a network Intrusion Detection System which can detect a wide variety of attacks reliably and efficiently when compared to the traditional network intrusion detection systems. Another advantage of our system is that it is very general and is easily customizable depending upon the specific requirements of individual networks. © 2011 Springer-Verlag.