摘要
用模糊集概念统计min-sup和minconf,并加入了第三约束要素:兴趣度,使min-sup和minconf通过数据信息本身的特性计算得到,规则可信度更高,避免了这两个值设置过高会异常漏检,设置过低无法检测异常的问题。根据这种思想设计了一种新的移动自组织网络入侵检测模型,把这个模型在网络仿真软件中对基于主机的数据进行了挖掘分析,用AODV协议实现了对模型的3种典型攻击。实验结果表明该模型对这些攻击的检测率平均达到90%以上。
The statistic of min- sup and minconf can be done by fuzzy set concept, at the same time the third key factor of confidence can be considered into it and min- sup and minconf can be gained by information characteristic itself. By this way, it avoids either the possibility of missing detection caused by too high value of min- sup and minconf or the possible disability to detect abnormality caused by too low value of the two variables. A new model of mobile self- constructed network intrusion checking was designed according to the thinking poimed above. The model was used in network- imitate software to do data mining based on the data of host computer. Furthermore, finished the testing of 3 representative attacks aimed to the model. The experiment results show that the average detection ratio of the model to these intrusion reached above 90 % .
出处
《计算机技术与发展》
2008年第7期135-138,共4页
Computer Technology and Development
基金
重庆市教育科研资助项目(040503)
关键词
数据挖掘
MANET
入侵检测
data mining
mobile ad- hoc networks
intrusion detection