摘  要： 在对网络入侵中边界模糊阈值进行确定的过程中，以往都是以入侵数据的单个异常特征为检测基础，没有考虑入侵后数据所表现的阶跃特征，出现网络入侵的判断不确定性，造成获取边界模糊阈值不准确的问题。文章提出一种基于阶跃模糊规则参数训练的网络入侵中边界模糊阈值的确定方法。该方法证明了网络遭到入侵后，数据整体带有阶跃性特征，依据这种阶跃性特征对入侵后的边界模糊阈值进行预测，塑造模糊系统的if-then规则，求出模糊系统的阈值输出，实现网络入侵中边界模糊阈值的确定。实验结果表明，该方法能较为准确地确定入侵模糊阈值，效果明显。

关键词： 网络入侵; 阶跃特征; 边界模糊阈值; 模糊if-then规则

中图分类号：TP301.6          文献标识码：A     文章编号：1006-8228（2021）08-37-05

The determination of boundary fuzzy threshold in network intrusion

Zhang Wumei

（ZhejiangTongji Vocational College of Science and Technology， Hangzhou， Zhejiang 310023， China）

Abstract： In the process of determining the boundary fuzzy threshold in network intrusion， it used to be based on the single abnormal feature of the intrusion data， without considering the step feature of the data after the intrusion， resulting in the uncertainty of network intrusion judgment， which leads to the inaccuracy of obtaining the boundary fuzzy threshold. In this paper， a method to determine the fuzzy threshold in network intrusion based on step fuzzy rule parameter training is proposed. This method proves that the whole data has a step feature after the network is invaded. According to this step feature， the fuzzy threshold after the invasion is predicted， the if-then rules of the fuzzy system are shaped， and the threshold from the output of the fuzzy system is obtained， so as to determine the boundary fuzzy threshold in the network invasion. The experiment results show that this method can determine the fuzzy threshold in intrusion more accurately， and the effect is obvious.

Key words： network intrusion; step feature; boundary fuzzy threshold; fuzzy if-then rule

0 引言

随着网络的广泛应用，网络入侵问题逐渐成为威胁网络安全的重点问题，而确定网络入侵中边界模糊阈值是解决网络入侵问题的重要途径。

文献[1]提出一种基于信号相位匹配原理的入侵阈值确定方法，采用该方法进行网络入侵信号的特征挖掘，其挖掘性能优于能量检测器，但算法对单频信号有效，在网络入侵信号存在的复杂信号环境中，确定效果不佳。文献[2]中引入DOA估计器，采用自相关预处理对时变信号进行跟踪，提高了入侵阈值确定的性能，但算法对信号的边缘入侵特征不能很好确定阈值，且不能满足对攻击信号实时检测的要求。网络入侵信号……