Abstract：Abstract: in metrological verification, we must consider the reliability of data, and it is also an indispensable factor in the uncertainty calculation of metrological standards. In multiple measurements, there are obvious differences between individual measurements and other measurements, which are called outliers or outliers. The existence of abnormal value will cover up the change rule of the measurement object and draw the wrong verification conclusion. This paper mainly introduces the common methods of distinguishing and dealing with outliers, and designs an intuitive, effective and fast outlier removal tool with Python language. This tool can save time and improve the accuracy of experimental measurement when used in daily verification work.
2020, Vol. 30 