摘要
本文描述了一种采用人工神经网络技术的高效异常入侵检测模型,对网络流量处理、神经网络的训练及其算法、神经网络的检测及其算法进行了详细的论述,利用Levenberg-Marquardt算法对传统BP算法进行改进,改进的BP算法较传统BP算法具有收敛速度快、正确检测率高的优点。
This paper describes in detail a Anomaly intrusion detection model based on neural networks,the processing of network flow and the training algorithm of neural networks.The Levenberg-Marquardt algorithm was taken to optimize traditional BP Neural Network,Compared with traditional BP algorithm,the new BP algorithm it brings up has higher speed in constringency and is more precise in detection.
出处
《网络安全技术与应用》
2007年第1期40-43,共4页
Network Security Technology & Application
基金
国家自然科学基金资助项目(69973016)
关键词
神经网络
异常检测
BP算法
neural networks
anomaly intrusion
the BP algorithm