基于遗传神经网络的入侵检测模型  

Intrusion detection method based on genetic algorithm and neural network

作　　者：刘道群[1] 孙庆和[2] 

机构地区：[1]重庆工学院,重庆400050 [2]四川通信科研规划设计有限责任公司,成都610041

出　　处：《激光杂志》2005年第6期73-74,共2页Laser Journal

摘　　要：这篇文章提出了一种基于遗传神经网络的入侵检测模型-进化神经网络入侵检测系统(ENNIDS),模型的核心模块利用遗传算法优化神经网络来实现,结合了误用检测和异常检测技术,并从理论上分析了该模型各个模块的功能和实现技术.我们在UCI机器学习数据库的入侵检测数据集上进行了实验,实验结果表明:该模型在检测正确率、误警率等方面能获得校好的性能。This paper proposes neural network and genetic algorithm based IDS model, Evolutionary Neural Network Intrusion Detection System （ENNIDS）. Neural network is optimized by IGA and then intrusion detection is performed by the optimized neural network, and misuse detection and abnormal detection are adopted. Experiments on the UCI dataset about intrusion detection results show that IGA has better performance than SGA on the best detection veracity and the fault alarming rate.

关 键 词：入侵检测 入侵检测模型 遗传算法 神经网络 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]