Application of fuzzy comprehensive evaluation in intrusion detection
Fei Zhao · Jisuanji gongcheng yu sheji · 2009
Based on the problem of representing intrusion features,sharp boundary,high false alarm rate and high miss alarm rate,the idea and model of the intrusion detection system with fuzzy comprehensive evaluation is presented.Then combined the methods of fuzzy mathematics with hierarchical analysis,a novel approach is provided to comprehensively evaluating the performance of intrusion detection systems.On one hand it uses synthetic weighting method of connecting advanced AHP with entropy method to configure or adjust the values of weight.On the other hand it uses non-fuzzy statistics method based on trapezium function and genetic algorithms to build the optimum membership function.This approach is applied to restrict the subjective factors that influence the evaluation and thus can improve the accuracy and reliability of the evaluation result.The effectiveness of this measurement is verified by an experiment.It’s turn to be compared with Snort being able to fall off high false alarm rate and high miss alarm rate.