数据挖掘中孤立点的分析研究在实践中应用  

Application of The Isolated Point Analysis Research in Data-Mining

作　　者：李强[1] 李振东[1] 

机构地区：[1]兰州市段家滩兰州商学院信息工程学院,兰州730000

出　　处：《微计算机应用》2006年第3期323-327,共5页Microcomputer Applications

摘　　要：介绍了孤立点的定义和三种挖掘算法,即基于统计的方法、基于距离的方法和基于偏离的方法。在这个基础上,尝试了利用孤立点检测方法对教务管理系统中积累的数据进行分析,并验证了基于距离和的孤立点检测算法的有效性。通过实验,结果分析表明:基于距离和的算法降低了检测过程对用户设置阈值的要求,在时间复杂度上,稍微优于循环嵌套算法。Outlier Mining is an important fact in data mining, it has special application and significance. This article introduced the definition of outlier and the algorithms for detecting outliers, that is statistical-based outlier detection, distance-based outlier detection, deviation-based outlier detection. On this base, this article try to use the detecting algorithms anglicizing the much data in the educational managing system, and an algorithm based on distance sum is proposed. The test result shows that the algorithm presented can reduce requirements of thresholds to users in the detection. This algorithm is a bit better than the nested-loop algorithm in the time consumption.

关 键 词：孤立点 数据挖掘 教务管理 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] O157.5[自动化与计算机技术—计算机科学与技术]