Intrusion Detection Techniques Based on Cutting-Based Real-Valued Negative Selection

Abstract

A novel technique for intrusion detection is proposed in this paper, that is based on cutting-based real-valued negative selection scheme. Different from the current typical techniques, the proposed one sets a much more strict and reasonable mechanism to generate and optimize the set of mutual detector. Firstly, the new generating detector must be necessary and it should not be detected by the current existing mutual ones. Besides, those detectors coinciding with the self-set should be cut and optimized into several qualified ones which have better detecting abilities. Experimental results demonstrate that the proposed technique has much higher detecting rates.