摘要
网络安全的问题日趋严重,入侵检测的研究是当今的研究热点。将数据挖掘和机器学习技术用于入侵检测是一个可行的方法。有很多算法用于入侵检测中,但有的是正确率比较低,也有的是学习或分类时间长,这些都限制了入侵检测系统在实际中的应用。文中提出了将粗糙集用于网络侦听的海量数据的属性约简,而后提出使用朴素贝叶斯进行分类预测。该方法的准确率高,而且时间性能好,适用于网络入侵检测的要求。
The technology of data minging and machine learning has been used in intrusion detection. The algorithm used in IDS needs that the accurate rate is high and the time of learming or classifying is short. Yet, lots of algorithms used in IDS cannot meet the needs which limit the use of IDS in pratice . In the paper,the naive hayes classifier based rough set reduction is proposed to use in IDS. The structure of naive hayes is simple,and learning corret efficiency and time efficiency is perfect. But it needs the independence of feature, which can be achieved by reduction based on rough set. It is fit for intrusion decahedron.
出处
《计算机技术与发展》
2006年第1期226-227,230,共3页
Computer Technology and Development
基金
安徽省教育厅自然基金资助项目(2002kj009)
关键词
入侵检测
朴素贝叶斯
粗糙集
属性约简
intrusion detection system
naive bayes
rough
set
festure reduction
