Web日志文件的异常数据挖掘算法及其应用

Algorithms for Mining Outlier Data on Web Log and Its Application

从数量化角度给出了异常数据的一般性定义，以Web服务器日志文件数据为依据，讨论了挖掘异常数据的方法和途径；给出了基于距离的单指标的离散统计法和综合统计法，并结合校园网作了实际的分析处理。结果表明，该方法是可行的。

This paper proposes the general definition of outlier data based on quantity, and then discusses the methods of mining outlier data on the basis of Web server log. It also proposes the discrete statistical and synthetical statistical scheme based on the criterion of distance and applies it to the campus network. The results indicate that the scheme is feasible.