基于测量信息论的小样本粗大误差处理研究

Processing of Gross Error in Small Samples Based on Measurement Information Theory

在小样本情况下,传统的基于统计原理的误差处理理论不能正确判定粗大误差。采用基于测量信息论的方法来判别测量数据中的粗大误差,将顺序统计量、秩和信息熵等概念引入误差处理理论用以判别小样本情况下的测量数据中是否存在粗大误差。分别采用基于测量信息论的方法和基于统计理论的误差处理方法,对实验数据进行处理和分析,表明基于测量信息论的粗大误差判别方法在小样本数据的粗大误差处理中具有优势。

In the case of small sample, the traditional methods of error processing based on statistical theory can not correctly distinguish the gross error. This paper uses a new method based on measurement information theory to process the gross error. Firstly, it introduces the concepts of order statistics, order and information entropy into the new method, then it uses this method to find the gross error in the measurement data. Three methods are used to process and analyze the experimental data ,and the result shows that the method based on measurement information theory is better at processing gross error in small samples than the methods based on statistical theory.