摘要
网络数据流在一段时间内会发生概念性变化,这可能会降低入侵检测的精度。针对网络数据流的这一特性,提出了一种能识别并适应概念飘移的基于滑动窗口的入侵检测方法,它能根据数据流的概念漂移的状况自动调整训练窗口并对检测模式进行及时的更新。
The data collected from network will change over a period of time in underlying concepts. This lowers the predictive precision of the detection. This paper proposes a method based on sliding window which can adapt to occurrence of concept drifting according to the characteristics of the network data stream. It can dynamically adjust the size of the training window and the detection model according to the current rate of concept drifting.
出处
《信息技术》
2009年第7期166-167,170,共3页
Information Technology
关键词
入侵检测
滑动窗口
概念漂移
数据挖掘
数据流
intrusion detection
sliding window
concept drifting
data mining
data stream
