摘要
讨论了利用神经网络设计识别用户异常行为的入侵检测系统的方案，即提取用户正常行为样本的特征来构造用户正常行为的特征轮廓；用神经网络扫描系统的审计迹得到的检测样本与用户特征轮廓进行比较，以两者的偏差作为证据，并结合证据理论来提高检测的正确率．
An intrusion detection system model based on the neural network and evidence theory is discussed. First we use a neural network to capture the user behavior pattern and to create the user normal behavior profile; then the deviant estimate of the detect samples is given by the neural network and the Dempster Shafer evidence theory is used to fuse the result derived from the neural network at different times, so that the abnormality in the user behavior can be detected more efficiently.
出处
《西安电子科技大学学报》
EI
CAS
CSCD
北大核心
1999年第5期667-670,共4页
Journal of Xidian University
基金
国家自然科学基金
关键词
神经网络
证据理论
入侵检测系统
计算机安全
neural network
evidence theory
computer networks security
intrusion detection system model
intrusion detection system
