摘要
提出了基于模糊Petri网的误用入侵检测方法,并将类似于神经网络的学习引入模糊Petri网,以调整攻击知识模型参数.理论分析表明,基于模糊Petri网的误用入侵检测系统具有更高的推理效率,能从环境中动态学习调整知识模型的相关参数,如阈值、权值、确信度.仿真结果表明,在大多数情况下,学习调整后的知识模型能够提高误用检测系统的检测率.
A method of misuse intrusion detection based on fuzzy Petri nets is proposed, and the learning ability similar to neural networks introduced into fuzzy Petri nets to adjust the parameters of attack knowledge model. Analysis indicated that, in the misuse detection system based on fuzzy Petri nets, the reasoning efficiency seemed to be improved, and the parameters such as threshold, weights and belief strength can be learned from the environment dynamically. Test results displayed that, under most circumstances, system detection rate was increased when the attack knowledge model was adjusted after learning.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2007年第4期312-317,共6页
Transactions of Beijing Institute of Technology
基金
国防科技基础研究基金资助项目(20021823)
关键词
误用检测
知识表示
模糊PETRI网
知识学习
misuse detection
knowledge representation
fuzzy Petri nets
knowledge learning
