摘要
数据缺失在社会经济研究、抽样调查、生物医药研究等诸多领域普遍存在,因而缺失数据的处理一直是国际统计学界热点讨论的课题之一。本文以哮喘临床试验为例构建缺失模型对缺失数据进行处理。通过介绍三种缺失机制,根据哮喘临床试验中真实情况模拟产生数据,在此基础上构建MAR模型,借助Win BUGS和R软件通过贝叶斯方法对模型中的参数进行估计。结果表明,在哮喘临床试验中基于缺失模型的方法对结论的敏感性分析效果显著。
Missing data is a main problem in many fields such as socio-economic research, sample surveys and the field of biomedical research and many other common. Therefore, coping with missing data has been an increasing important issue in the discussion of international statistic. In this paper, we built a model for coping with the missing data from a asthma clinical trials. We introduced three kinds of missing mechanisms to analyze the character of missing data in different missing mechanism. The data in this paper were generated from real situation simulation with R,based on which MAR model were built. The results show that, in asthma clinical trials,the method we use is more preferable to assess the sensitivity of the conclusions.
出处
《价值工程》
2015年第31期187-189,共3页
Value Engineering
