江西科学Issue(1)：47-51,5.DOI:10.13990/j.issn1001-3679.2016.01.012
基于WTLS的部分最小二乘粗差探测法
Partly Least Square Outlier Detection Method Based on WTLS
摘要
Abstract
For the problem of outliers in EIV models,a new outlier detection method of EIV model is introduced. Assuming a set of observation vector without outliers is adjusted orderly,then using the weighted total least squares after deformation. To detect outliers in observations,the estimation of unit weight variance can be used to classify observations. The numerical example reveals that the new method can effectively locate the outliers in observations when the observations have the same weight.关键词
EIV模型/粗差探测/粗差/加权整体最小二乘/单位权方差Key words
EIV models/outlier detection/outliers/weighted total least squares/unit weight variance分类
天文与地球科学引用本文复制引用
王铮尧,鲁铁定,焦亦詹..基于WTLS的部分最小二乘粗差探测法[J].江西科学,2016,(1):47-51,5.基金项目
国家自然科学基金（41204003,41374007,41464001) （）
江西省科技落地计划项目（KJLD12077) （）
江西省教育厅科技项目（GJJ13457) （）
中国博士后基金项目（94773) （）
江西省中青年教师发展计划访问学者专项（2012-132)。 （）
