【作者】 赵云丰；

【导师】 刘万军；

【作者基本信息】 辽宁工程技术大学 ， 计算机应用技术， 2005， 硕士

【摘要】 入侵检测技术的进一步发展,给网络安全的研究带来了新的途径,入侵检测系统已成为必不可少的重要手段。为了提高入侵检测系统的检测能力,本文提出了一个ID模型。在模型中引入了神经网络NN技术,提出了一种改进训练算法,并提出了一些提高入侵检测系统检测能力的方案,探讨了流量分析、网络阻断、IP追踪和高速地址匹配等技术,最后对模型的核心组成部分进行了分析、设计,并进行了相关实验。本文是把神经网络应用于入侵检测的一次尝试,它摒弃常规的基于行为的ID模式,采用了更先进的、模拟人脑神经网络系统的非线性工作模式。神经网络还可结合专家系统、数据挖掘等技术,神经网络在ID中的应用有着巨大的理论及实践意义。更多还原

【Abstract】 With the further development of intrusion detection technology, it provides a new way to study the network security. IDS has become an important and indispensable method. To improve the ability of Intrusion Detection System,this text has proposed a kind of ID model, Neural Network Intrusion Detection Model based on neural network. In the model, we have introduced the NN, and propose a kind of improvements train algorithm. And we are going to discuss some problems, such as intruder tracing, data flow analysis, and address matching, and network blocking. We analyze, design and realized the key part of the model at last, and carried on the relevant experiment.This article is an attempt which we applies NN to ID, we abandon the routine ID model based on behavior and adopt more advanced one, the non-linear work pattern of neural network system of people of simulation. NN can also combine expert system, data mining, and so on. The NN application in ID has enormous theory and practice meaning.更多还原