摘要
提出了神经网络驱动模糊推理的入侵检测方法,利用神经网络的学习能力,对不清楚规则的复杂系统的输入输出特性进行适当的非线性划分,自动形成规则集和相应的隶属关系,克服了在多维空间上经验性的确定隶属函数的困难。对于神经网络的训练数据,采用人工数据,克服了神经网络监督学习和获取已知输出的训练数据的困难。试验证明,这种技术具有很好的灵敏度和鲁棒性,而且,能够检测出未知的入侵行为。
This paper describes a novel intrusion detection method based on fuzzy reasoning drived by neural network (NN). In order to overcome the difficulty of specifying the membership functions of rules depending on experiences of experts in multi-dimension space, neural network is introduced to distinguish non-linearly input/output characteristics of complex system and to generate rule sets and membership functions automatically. The NNs in this experiment are trained using data generated artificially, eliminating both problems, which are the facts that A BP NN is initialized randomly and must undergo' supervised learning' before being used as a detector and that obtaining training data with knowledge of the desired output for each input vector. The technique demonstrated in this experiment appears to be sensitive and robust, moreover, which is able to detect unknown attack and plays down false alarms and missing alarms.
出处
《计算机工程》
CAS
CSCD
北大核心
2003年第19期133-135,共3页
Computer Engineering
基金
国家"863"计划基金资助项目(863-301-05-03)
关键词
神经网络
模糊推理
入侵检测
Neural network
Fuzzy reasoning
Intrusion detection
