一种基于神经网络的入侵检测系统研究

提出在入侵检测系统模型设计中引入神经网络技术,建立了一个基于神经网络的入侵检测系统模型。针对传统BP算法存在的一些固有缺点,提出增加动量项与自适应调节学习速率相结合的改进算法,一定程度上克服了BP神经网络存在的问题。实验结果表明,基于改进的BP神经网络的入侵检测方法具有良好的检测性能。

The Research of Intrusion Detection System based on Neural Network

In this paper,we propose the application of neural network technology to the design of the Intrusion Detection System,and builds a model of intrusion detection system based on neural network. To the inherent shortcomings of the traditional BP algorithm,here we use the improved BP algorithm which combines adding momentum item with the adaptively controlling learning rate,in a certain degree,can overcome these shortcomings. The evaluation results show that the performance of the Intrusion Detection method based on improved BP algorithm is effective.