中文摘要：

对于缺失数据的处理,通常会采用直接删除缺失数据或者插补数据的方法来得到完整的数据集。多重插补法是对缺失数据集进行多次插补,并且提高数据稳定性的一种较好的缺失数据处理方法。基于多重插补法分析公路安全数据,进而对完整的数据集进行方差估计的比较。经过具体的计算和分析可以得到结论,多重插补法可以得到信息量相对完整的数据集,利用这样的数据集得到的分析结果更加可靠。

英文摘要：

The method of censoring missing values or imputing missing values to obtain a complete data set is often used to deal with the missing data. The multiple imputation method that can impute the missing data set repeatedly and improve the data stabilization is preferable. This paper compares the variance estimation among the complete data set. After calculations and analysis, it can come to the conclusion that multiple imputation method can get more complete and effective information which can make the analysis results more reliable.