Database Intrusion Detection Using Adaptive Resonance Network Theory Model

In the current automation world, every organizations starting from education to industry and also in research organization are maintaining database and several security breaches are found in these databases. Traditional database security mechanism cannot handle the malicious access toward database. Although, various researches have been done in database intrusion detection, but most of the researches are limited in efficiency and accuracy. Inspired by human cognitive system, we present a database intrusion detection system using Adaptive Resonance Theory which is accompanied with some data mining techniques for preprocessing data. The proposed model can learn easily and cope up with the dynamic environment which entitles the system to detect both known and unknown patterns accurately with low false positive cost. The calculated simulation result shows that the database intrusion detection based on Adaptive Resonance Theory can accelerate the detection process with higher accuracy as compared to Self Organizing Map and Radial basis functional neural network.