基于遗传神经网络的入侵检测系统ONE-CLASS分类器设计  

Design of One-class Classification for Intrusion Detection Systems Based on Genetic Neural Networks

作　　者：戴月[1] 陈波[2] 吴坚[2] 

机构地区：[1]滁州职业技术学院,安徽滁州239000 [2]西南科技大学,四川绵阳621010

出　　处：《微计算机信息》2011年第7期194-195,71,共3页Control & Automation

摘　　要：为适应高速网络中的数据处理速度,设计了将识别出的正常数据抛弃的入侵检测系统one-class分类器。检测模块采用GA与BP相结合的智能算法。该算法利用神经网络自身具有并行性、鲁棒性等特点,可以大大减少分类器的计算时间。In order to adapt the data processing speed on the high-speed network, one-class classification for intrusion detection systems based on genetic neural networks was designed to abandon the normal data be identified . Making use of neural networks oneself characteristics of parallelism and robustness, a detection method that joined the genetic algorithm and BP algorithm together was proposed to reduce calculation time of one-class classification consumedly.

关 键 词：入侵检测系统分类器 遗传算法 BP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]