基于狄克松准则剔除水文数据异常值

Removing Abnormal Values of Hydrological Data Based on Dixon Criterion

为保障水文数据准确性,以较大波浪情况下水位观测数据为例,提出狄克松准则异常值检查方法。首先,对数据进行从小到大排列,并对两端和被怀疑数据进行统计分析,然后根据信度查得临界值,将得到数据进行对比,剔除异常值,与3准则进行对比分析,检查异常值准确度优于3准则,最后通过不确定度来验证观测数据精度。

In order to ensure the accuracy of hydrological data,a method for checking the abnormal values of Dickson criterion is proposed in this paper,taking the observed data of water level under large waves as an example.First,the data is arranged from small to large,and statistical analysis is performed on both ends and suspected data.Then,depending on the reliability,the critical value is found,and the data is compared to remove the outliers and compared with 3σCriteria are compared and analyzed,and the accuracy of checking outliers is better than 3σGuidelines.Finally,the uncertainty is used to verify the accuracy of the observed data.