Improved Hybrid Algorithms using Data Mining Techniques for Intrusion Detection System

The tremendous growth of the usage of computers over network and development in application running on a various platform captures the attention towards network security. This paradigm exploits security vulnerabilities on all computer systems that are  technically difficult and expensive to be solved.  As a result, Intrusion Detection Systems (IDSs), are widely used to detect and stop network intrusions [1,2]. An intrusion detection system watches networked devices & searches for malicious behaviors in kinds of pattern in the audit stream [13]. One main conflict in intrusion detection [3] is that we have to find out the hidden attacks from a large quantity of routine communication activities. The security of our computer systems & data is at continual risks due to the extensive growth of the internet & increasing availability of tools & tricks for intruding & attacking networks have made intrusion detection to become a critical component of network administration. In traditional intrusion detection systems,  the network administrator needs to spend much time in analyzing the data and also generates high false alarm rates.  But data mining based intrusion detection systems are more precise, require less manual processing time and input from human experts.

G.V.Nadiammai