摘要
入侵检测系统是保证网络信息安全的有力手段,文中提出一种结合决策树和神经网络的入侵检测系统框架。决策树分类方法把数据集划分为正常数据和入侵数据,并作为训练集分别用神经网络进行训练,改善了系统的检测精度并提高了对未知数据的检测能力。离线训练后的系统可以实现网络数据的实时检测,通过实验证明了此系统很好的检测效果和自适应能力。
Intrusion detection system is an efficient method for information security. An intrusion detection system framework based on decision tree and neural network is proposed in this paper. Dataset can be labeled normal or intrusion based on decision tree, which are transferred to neural network as training dataset. After training of the neural network, IDS improved its accuracy and ability to detect new intrusion. IDS can detect intrusion online efficiently and adaptively in our experiment.
出处
《燕山大学学报》
CAS
2010年第1期85-89,共5页
Journal of Yanshan University
基金
河北省科技支撑计划资助项目(072135218)
关键词
入侵检测
决策树
神经网络
intrusion detection
decision tree
neural network
