摘要
为克服目前入侵检测技术检测反应速度慢、误检率和漏检率较高等问题,研究了加权移动窗口这种数据挖掘方法。首先对现有的移动窗口算法MFI-TransSW和Moment进行了认知与分解,指出现有算法的缺陷,提出了加权移动窗口的详细算法,自动调整训练窗口,并对检测模式进行及时的更新;在此基础上建立了基于加权移动窗口的入侵检测系统模型。最后实例检测和结果分析表明,在不同窗口大小、不同最小支持度、数据集增大时该算法执行时间均优于其他算法。
To overcome the current intrusion detection question that detecting response was slow,and false detection rate and false negative rates were relatively high and so on,studied the weighted moving window of this data mining method.Firstly,recognized and decomposed the existing mobile window algorithms.It pointed out that the shortcomings of existing algorithms,presented the detailed algorithm of the weighted moving window,adjusted the training window automatically,and updated timely the patterns of detection.Based on the weighted moving window algorithm,constructed a weighted moving window-based intrusion detection system model.Finally,an example testing and result analysis shows that the execution time of the algorithm in different window sizes,different minimum support and the increasing data sets is better than other algorithms.
出处
《计算机应用研究》
CSCD
北大核心
2010年第7期2643-2646,共4页
Application Research of Computers
关键词
入侵检测系统模型
数据挖掘
移动窗口算法
加权移动窗口
网络数据流
频繁集
支持度
intrusion detection system model
data mining
moving window algorithm
weighted moving window
network data flow
frequent sets
degree of support
