摘要
异常数据检测是数据挖掘研究的热点之一。本文在对现有异常点检测算法分析的基础上,提出了一种基于属性的异常点检测算法。简要地介绍了异常检测的现状,对基于属性的异常检测算法进行了详细分析,包括算法设计基础、算法描述、复杂度分析等。并通过与基于距离的异常点检测算法进行实验比较,表明了算法的优越性。
Outlier data detection is an important part of data mining. It is a hotspot in data mining researc. Based on the analysis of the exsiting algorithms of outlier data detection, this paper put forward a new outlier detection algo- rithm based on attribute. We introduce the status quo of outlier detection briefly, and analyze the algorithm of outlier detection based on attribute particularly. This paper shows the design basis of the new algorithm, the depiction of the new algorithm and the analysis of the complexity of the new algorithm and so on. Compared with another algorithm based on distance by experiment, the new algorithm has an obvious superiority in detection precision and time con- sumption.
出处
《计算机科学》
CSCD
北大核心
2005年第5期164-166,共3页
Computer Science
