结合影响分析的抗隐差型Bayes粗差探测方法
Bayes Unmasking Method to Detection of Gross Errors Together with Influence Analysis
-
摘要: 通过分析隐差现象产生的原因,将基于观测误差的后验概率法与基于Kullback-Leiber距离的影响分析相结合,并利用逐个搜索法的思想,给出了一种粗差探测的抗隐差型Bayes方法。实验结果表明,采用该方法进行粗差探测切实可行,不但有效地发现了被掩盖的粗差,而且运算简单,效率较高。Abstract: By analyzing the reason of masking and combining the method of posterior probability of observation error and the influence of Kullback-Leiber divergence,a Bayes unmasking method for gross errors detection is proposed based on the idea of one-by-one searching.The experimental results show that this method is successful for gross errors detection.The method not only detects the gross errors of masking effectively and has a bstter result,but also has a simple computational process and higher efficiency.
下载:
