基于神经网络的网络入侵检测系统设计

Design of Network Intrusion Detection System Based Neural Network

把神经网络作为异常检测系统的统计分析部分的一种替代方法,用来识别系统用户的典型特征,对用户既定行为的重大变化进行鉴别。将模式匹配与人工神经网络技术结合在一起,构成一个以已知的入侵规则为基础、可扩展的动态入侵事件检测系统。自适应的进行特征提取与异常检测,实现高效的入侵检测及防御。用神经网络来过滤出接收数据当中的可疑事件,并把这种事件转交给系统作进一步的处理。这种结构可以通过减少系统的开销和IDS误报率来提高监测系统的效用。

The neural network is used as an alternative method of statistical analysis in anomaly detection system, which is used to identify the typical profile of users of the system and identifies significant changes of establishing user behavior. Integrating pattern matching with artificial neural network technology constitutes dynamic intrusion detection system which is based on known rules of the invasion. Adaptive feature collection and anomaly detection can achieve efficient intrusion detection and defense, which using neural networks to filter out suspicious data and transmit this data to the system for further treatment, This architecture can reduce system overhead and IDS＇s false alarm rate to improving the efficiency of the detection system.