原型观测数据离群点的检验  

Detection on outlier of prototype observation data

作　　者：马宇燕[1] 杨建宇[1] 

机构地区：[1]北京华可实工程技术有限公司,北京100025

出　　处：《水利水电技术》2010年第6期19-20,24,共3页Water Resources and Hydropower Engineering

摘　　要：在水工建筑物的原型观测数据处理工作中,经常会遇到一些异常值(离群点)。本文通过深入分析,判断离群点产生的性质和原因,从数量上揭示数据的变化规律,以动态的角度表现某一工况条件下,各变量之间的内在关系,从而达到认识水工建筑物运行状态的目的。During the prototype observation of hydraulic structures, some abnormal values （outlier） often occurs in the treatment of the observed data. The causation and character of the occurrence of the outlier are judged herein through a detailed analysis ; in which the variable law of the data is quantitatively revealed and the inherent relationship among all the variables under a certain operation condition is dynamically presented as well, so as to acquaint the operation condition of the hydraulic structures concemed.

关 键 词：时间序列 离群点 6倍标准差 原型观测 水工建筑物 

分 类 号：TV698.1[水利工程—水利水电工程]