面向入侵检测的基于多目标遗传算法的特征选择

Feature Selection Using Multi-Objective Genetic Algorithms for Intrusion Detection

针对刻画网络行为的特征集中存在着不相关或冗余特征,从而导致入侵检测性能下降的问题,本文提出了一种基于多目标遗传算法的特征选择方法,将入侵检测中的特征选择问题视为多目标优化问题来处理。实验结果表明,该方法能够实现检测精度与检测算法复杂性的均衡优化,在显著提高检测算法效率的同时,检测精度也有所提高。

A feature selection method using multi-objective genetic algorithms is proposed to solve the problem of performance degradation of the intrusion detection, which results from the existence of irrelevant or redundant features among the feature set representing the network behavior, The method views the feature selection for intrusion detection as multi-objective optimization problem. The experimental results manifest that the best detection accuracy/complexity trade-off can be achieved. The detection accuracy is better, while the detection algorithm efficiency is improved remarkably.