关键词: 计算机系统结构 , 入侵检测 , 学习Petri网 , 神经网络

Abstract: A method of intrusion detection based on neural network(NN) has flaws of slower learning speed, hardness in converging and deficiency of classifier capability. The learning Petri nets(LPN) were adopted to construct the method of network intrusion detection. LPN is superior to NN in the re?alization of nonlinear and discontinuous functions. The test result indicates that the classifier based on LPN has better recognizing precision and faster learning speed compared with the classifier based on the same structure NN.

Key words: computer system architecture , intrusion detection , learning Petri nets , neural network