【摘要】 针对神经网络检测器本身的网络结构和算法进行改造可获得好的性能,但无法从根本上解决误报率和漏报率等问题,通过对程序行为的深入研究,对程序行为进行动态建模,提出了一个应用BP神经网络检测器针对程序行为异常的入侵检测模型,从而更准确地发现程序行为的异常。通过Apache服务器为例论证其可行性。更多还原

【Abstract】 Neural network can be used in anomaly detection. In order to improve the traditional IDS (intrusion detection system) performance, we often had to change the network structure itself and detection algorism. And because intrusion techniques are most changeful and unpredictable, we can not always use the fixed detection techniques to catch exactly all possible intrusions. It is therefore important to investigate novel detection methods and IDS models. In this paper, a model of intrusion detection system based on BP neural network is proposed through analyzing characteristic of program behavior. Some details and issues on the design and implementation of the model are discussed and an experiment is also given.更多还原