数据校正源码。1、粗大误差识别与剔除功能有多种准则(莱依特、肖维勒、罗蒙诺夫斯基、格拉布斯、狄克逊)，不同准则算法不一样。在一个剔除周期内，对同一变量的n的样本进行粗大误差识别和剔除，样本个数n大于5，小于等于102.2、数据协调功能再粗大误差被剔除的基础上，在满足约束条件下求出无偏最小方差估计解，此解即为数据协调值。-Data correction source. 1 gross error identification and elimination features a variety of criteria (Levin Eternalchem, Xiaowei Le, Romon Malinowski, Grubbs, Dixon), different criteria are not the same algorithm. The coarse error recognition and rejection, a rejection period, the sample of the same variable n and the number of samples n is greater than 5 and less than or equal to 102.2, the data coordination function and then coarse error basis been removed, to meet the constraint condition obtained unbiased minimum variance estimation solution, this solution is the coordination of data values.