摘要
在对复杂网络数据流的入侵数据进行检测时,由于网络属性的复杂性,使得数据流的类型和流向在时间上存在着复杂变化,采用传统方法进行检测时,复杂性造成数据不规则性更强,使得数据的特征与属性参数频繁发生变化产生众多数据特征,导致检测准确度低、效率差的问题。提出改进WSW-Imp算法的复杂网络数据流的入侵数据检测方法,融合于倾斜时间窗口策略,利用高精确度的时间粒度来保存数据流内实时的频繁模式信息,在以数据流特征属性划分的信息增益为标准将全部数据流的特征划分为多个离散区间,同时采用线索化二叉排序树结构组建增量模糊决策树,进而有效的完成了对复杂网络数据流的入侵数据检测。实验结果证明,改进WSW-Imp算法的复杂网络数据流的入侵数据检测准确度高。
In the paper, a method for intrusion data detection of complex network data flow was proposed based on improved WSW - Imp algorithm, and a tilted - time window strategy was fused to the method. The high precision time granularity was used to save the real - time frequent pattern information in data flow. Then, taking the information gain of data flow characteristic attribute partition as the standard, all characteristics of the data flow were divided into multiple discrete intervals, at the same time, threaded binary search tree structure was used to construct increment fuzzy decision tree, thus to effectively complete the intrusion data detection of complex network data flow. The experimental results prove that the accuracy of intrusion data detection of complex network data flow is high by using improved WSW- Imp algorithm.
出处
《计算机仿真》
CSCD
北大核心
2015年第12期272-275,共4页
Computer Simulation
基金
<计算机应用基础>教学系统的设计与实现(JYYB201404)
关键词
数据流
数据检测
数据流检测
Data flow
Data detection
Data flow detection
