Parametric transformed Fay-Herriot model for small area estimation
Parametric transformed Fay-Herriot model for small area estimation
复制标题
用于小面积估计的参数变换 Fay-Herriot 模型
DOI：
10.1016/j.jmva.2015.04.001
复制
发表时间：
2015
影响因子：
1.6
通讯作者：
S. Sugasawa and T. Kubokawa
中科院分区：
文献类型：
--
作者：
Sugasawa;S and Kubokawa;T.;Y. Ikeda and T. Kubokawa;S. Sugasawa and T. Kubokawa
本文在分析收入、税收、收成和产量等实证数据的基础上，提出了小面积估计中的参数变换Fay-Herriot模型。当双幂变换被用作参数变换时，我们给出了变换参数、回归系数和方差分量的相合估计。经验最佳线性无偏预测，插入这些一致的估计建议，和他们的均方误差（MSE）的渐近评估。通过参数自助法，给出了均方误差的二阶无偏估计。最后，通过仿真和实证研究，所建议的程序的性能进行了研究。
Motivated from analysis of positive data such as income, revenue, harvests and production, the paper suggests the parametric transformed Fay–Herriot model in small-area estimation. When the dual power transformation is used as the parametric transformation, we provide consistent estimators of the transformation parameter, the regression coefficients and the variance component. The empirical best linear unbiased predictors which plug in those consistent estimators are suggested, and their mean squared errors (MSE) are asymptotically evaluated. A second-order unbiased estimator of the MSE is also given through the parametric bootstrap. Finally, performances of the suggested procedures are investigated through simulation and empirical studies.