摘 要在当前社会的各个领域研究调查中,因为各种已知或未知因素的影响,常常会导致缺失数据这种情况的出现.缺失数据的存在,不仅会增加研究专家们分析数据时的复杂程度和难度,而且还会造成分析结果的偏差,从而降低工作者的工作效率,因此考虑怎么消除或者尽可能的减小这些缺失数据的影响就变的更加重要.本文首先是用几种填补法对缺失的数据进行填补,构造一个完整的数据集,然后再利用SPSS软件对填补后的数据集进行相应的统计分析. 本篇毕业论文主要针对缺失数据处理的方法进行分析研究,主要方法有均值填补法(Mean),期望值最大化填补法(EM),回归填补法(Regression)和多重填补法(MI).19068
关键字:数据缺失;回归填补法;期望值最大化填补法,均值填补法
The Estimation And Application of Missing Data
Abstract In the current study investigated in all areas of society, because of the effect of all kinds factors that known or unknown, it always leads to missing data for this situation. The presence of the missing data not only increase the complexity and difficulty that research experts analyze the data, but also can lead to deviation of analytical results, thereby reducing the working efficiency of the workers, and as much as possible to consider how to eliminate or reduce the impact of these missing data has become more important. Firstly, it need several filling methods to fill the missing datas , constructing a complete data set, and then using the SPSS software to fill the corresponding statistical analysis for the data set. Processing method of this paper is aimed at missing data analysis, the main methods are mean imputation methods(MEAN), expectation maximization method (EM), regression imputation (Regression) and multiple method (MI).
Keywords: Missing Data; Regression ; EM; Mean
目 录