自适应动态入侵检测模型的研究

Design and Implementation of Adaptive Dynamic Intrusion Detection Model

提出了将模糊控制和神经网络用于入侵检测的新方案.在该模型中神经网络训练模块可以不断地从模糊控制模块中获得攻击数据信息和正常数据信息,并根据这些信息进行自适应调整,更新权值和阈值,使整个智能入侵检测过程完全成为一个在实际应用中动态自适应的过程.实验结果表明,这种方案具有很高的准确检测率,对检测未知攻击具有较好的性能.

This paper puts forward a new program of employing fuzzy control technology and neural network theory to detect intrusion. In this model, neural network practice model can constantly obtain charge data information and normal data information, upon which the model can make self-adjustments and renew weight value and threshoed value, thus making the whole intrusion detection process a dynamic and self-adjusting one in real use. Experimental results reveal that this program has a high rate of correct detection and a better function in detecting unknown charge.