统计报表中异常值的识别与缺失值的预测

The Identification of Outliers and the Forecast of Missing Values in the Statistic Report

本文通过两个统计年报表的例子验证了统计报表中的数据近似服从对数正态分布,并利用这一特点建立了对数线性模型,对统计报表中的异常值进行识别和对缺失值作出预测。

Through the empirical study on two annual statistic reports, the paper shows that the data in the statistic report follows log-normal distribution approximately. In terms of this propety, the paper builds up a log-normal model, and applies it to identify the outliers and forecast the missing values in the statistic report.