大地测量与地球动力学2012，Vol.32Issue(2)：51-54,59,5.
基于Kullback-Leiber距离粗差探测的Bayes方法
BAYESIAN APPROACH TO DETECTION OF GROSS ERRORS BASED ON DIVERGENCE OF KULLBACK-LEIBER
摘要
Abstract
With the view of influence analysis, a new Bayesian method for gross errors detection based on divergence of Kullback-Leiber is proposed. Under the condition of unequal weight and independent observations, a comparison among the data deletion model, the variance inflation model and the mean shift model based on certain prior distribution is made, and the computational formula of divergence of Kullback-Leiber about the three models is given and a judge rule of gross error detection is established. Finally, the method is used for gross error detection in triangulateration network and a good result is obtained.关键词
Bayes方法/影响分析/Kullback-Leiber距离/粗差探测/方差膨胀模型Key words
Bayesian method/ influence analysis/ Kullback-Leiber divergence/ gross errors detection/ variance in-flation model分类
天文与地球科学引用本文复制引用
王延停,归庆明,张倩倩..基于Kullback-Leiber距离粗差探测的Bayes方法[J].大地测量与地球动力学,2012,32(2):51-54,59,5.基金项目
国家自然科学基金(40974009) （40974009）
郑州市科技计划攻关项目(0910SGYG21198) （0910SGYG21198）
