基于反馈学习的网络攻击过滤方法研究

摘要

This paper proposed a network attack filtering method based on feedback learning. The feedback train-ing was used to delect the wild attacks, and the error drive was used for training to complete the accurate classification for wild attacks. As the feedback processed, the changing in attack types can be captured by feedback learning, thus the network classification performances were gradually improved. The experimental results prove that the method can quickly and accurately complete the filtering of bad attacks.%研究网络不良攻击的过滤问题,提高网络抗攻击能力,可保证网络安全.针对当前网络中攻击识别过滤问题,当过滤网络攻击时,只能够明确归属的攻击类型,当网络攻击不能明确归属特定类型,会产生攻击分类效果下降,产生野攻击,导致不能有效过滤的问题.为了解决上述问题,提出了一种反馈学习的网络攻击过滤方法,通过反馈训练对野攻击进行检测,利用错误驱动进行训练,完成对野攻击的准确分类.随着反馈的进行,反馈学习能捕捉到网络中攻击种类的变化,分类性能逐步提高.实验证明,改进方法能快速准确的完成网络不良攻击的过滤,取得了明显的效果.