基于组合神经网络的网络入侵检测  

Intrusion Detection based on Combined Neural Networks

作　　者：欧阳陈华[1] 康江林[1] 旷海兰[1] 李浪[1] 

机构地区：[1]衡阳师范学院计算机系,衡阳421008

出　　处：《科学技术与工程》2006年第13期1932-1934,1942,共4页Science Technology and Engineering

基　　金：衡阳师范学院科学基金(2005A05);湖南省教育厅科研项目(05B055)资助

摘　　要：提出了一种构造组合神经网络的方法。该方法包括两个步骤,先在粗糙集约简的基础上确定单个神经网络的结构,以及在组合时要包含神经网络的个数;然后将这些分类器根据属性重要性组合起来。用该组合神经网络进行网络入侵检测,实验结果表明该组合方法是一种有效的入侵检测方法。Some methods are put forward to construct combined neural networks. Those methods consist of two steps, first of all, on the basic of reduct of rough set to make sure the structure of single Neural Network, and the unity of every included neural network when it combined. Then, according to the importance of quality, the classifier will be reorganized. The combined neural networks will be used to detect the network intrusion, afterwards the experiment result shows the methods are valid.

关 键 词：粗糙集 组合神经网络 入侵检测 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]