样本选择模型的似然估计与两步估计  

LIKELIHOOD ESTIMATION AND TWO-STEP ESTIMATION OF SAMPLE SELECTION MODEL

作　　者：张磊[1] 王彤[1] 王琳娜[1] 

机构地区：[1]山西医科大学公共卫生学院流行病与卫生统计教研室,太原030001

出　　处：《现代预防医学》2007年第9期1607-1609,共3页Modern Preventive Medicine

基　　金：山西省高校青年学术带头人基金项目(晋教科2004-13号)

摘　　要：[目的]介绍样本选择模型及其估计方法。[方法]应用样本选择模型对模拟数据进行分析,并与传统线性回归模型进行比较。[结果]对于特定条件下的缺失数据,样本选择模型比传统的线性回归模型更合理,并可以提高模型参数估计的效率。[结论]样本选择模型在医学研究中拥有广泛的应用前景。[ Objective] To introduce sample selection model and its estimation method. [ Methods] The missing data were analysed with sample selection model and compared with traditional linear regression model. [ Results] When missing data had special structure, sample selection model was more sensible and efficient compared with traditional linear regression model. [Conclusion] The sample selection model has good applicability to medical research.

关 键 词：样本选择模型 似然估计 两步估计 缺失数据 

分 类 号：R195.1[医药卫生—卫生统计学]