Feature Optimization Based on Artificial Fish-Swarm Algorithm in Intrusion Detections
Tao Liu, Ailing Qi, Hou Yuan-bin, Chang Xin-tan · 2009
A method of optimization and simplification to network feature using Artificial Fish-swarm Algorithm in intrusion detection is proposed in this paper for solving problems of more features and slower computing speed. This method established mathematic model aimed at achieving higher detection rate and lower false positive rate, and obtaining optimal feature attributes through iterative method by using an optimization policy on the basis of "PREY, SWARM and FOLLOW" operators. 41 features are optimized and simplified by adopting this method. 31% feature attributes are achieved, which can completely reflect intrusion feature. The experimental results show that using feature attributes after optimization and simplification can shorten 40% work time in intrusion detection.