一种神经网络和模式匹配相结合的入侵检测系统

构造一个基于BP神经网络和模式匹配技术的网络入侵检测系统模型.神经网络和模式匹配在检测类型上是互补的，BP神经网络需要数值化的输入，适合检测基于网络数据包流量特性的入侵行为；模式匹配技术是通过在数据包中搜索特征字符串来检测入侵的.将网络数据包分析处理后分别按神经网络和模式匹配检测模块的输入数据格式生成输入实例，再分别予以检测.该系统可以检测出已知的入侵行为，同时具有一定的对未知入侵的检测能力.实验证明两者结合起来构成的检测系统性能更全面.

A neural network and pattern matching-based intrusion detection system

A neural network and pattern matching-based net intrusion detection system is designed.In the capacity to check intrusion methods,neural network and pattern matching can complement each other.BP neural network needs number input and is fit for detecting the intrusion based on the character of data packets flux of net.Pattern matching detects the intrusion by searching specific character string in the data packet.After dealing with the net packet,the input examples are generated according to the input formats of neural network and pattern matching respectively,and then intrusion is detected.The system can detect the known intrusion and has some ability to detect the unknown intrusion.Experimental results testify that the intrusion detect system combined the two technologies has better detecting capacity.