研究了比例Bootstrap,介绍其与通常的Bootstrap方法的差异;对简单随机抽样,在均匀回答的假定下,就缺失数据采取回归插补和比率插补的情形,证明了由比例Bootstrap方法得到的方差估计具有相合性;最后,通过模拟验证了理论结果.
This paper studies the proportional Bootstrap method on missing data, and compares it with the usual Bootstrap method. Under the simple random sampling and uniform response, using regression imputation and ratio imputation for missing data, the consistency of the variance estimators obtained by proportional Bootstrap is proved. Simulation results give a verification of our theoretical conclusions.