Associate Professor,

Active worms has been major security threat to the Internet. This is due to the ability of active worms to propagate in an automated fashion as they continuously compromise computers on the internet. Active worms develop gradually during their propagation and thus pose a great challenges to protect from danger against them. In this paper we study a new class of active worm, referred as a camouflaging worm. The camouflaging worms is different from the normal traditional worms because of it has the ability to intelligently manipulates its scan traffic volume over time. So it is very difficult to detect the camouflaging worm by using normal worm detection technique. We propose a new worm detection technique based on Machine Learning Technique. This approach captures dynamic program behavior to provide accurate and efficient detection against both seen and unseen worms. This particular software is developed based on the data set given by DARPA. The data set contains lakhs of transactions that include all attacks observed by the DARPA. Based on the transaction behavior we made software. Our experimental results clearly demonstrate the effectiveness of our approach to detect new worms in terms of very high detection rate and a low false positive rate. This process not only detect camouflaging worms also detects the denial-of-service (DOS) attacks, probe attacks and etc