Event Details
Title: The empirical best predictor in the two-fold nested error regression model
Speaker: Domingo Morales (Universidad Miguel Hernández)
Abstract: We propose to estimate nonlinear small area population parameters by using the empirical Best predictor based on a two-fold nested error model. The introduced statistical methodology is applicable to poverty indicators and more generally to separable nonlinear parameters paper. We use a parametric bootstrap method to estimate the mean squared error of the empirical best estimators. We also study small sample properties of these estimators by model-based simulation studies. The simulations show reductions in mean squared error relative to direct area-specific estimators and other estimators obtained by “simulated” censuses. We also apply the proposed method to estimate poverty incidences, poverty gaps, and the corresponding mean squared errors in Spanish provinces by gender. For the Spanish data, results show a significant reduction in coefficient of variation of the proposed empirical best estimators over direct estimators.
27 November 2015, 12:30, 1st floor Seminar room, IEMath-GR
Speaker: Domingo Morales (Universidad Miguel Hernández)
Abstract: We propose to estimate nonlinear small area population parameters by using the empirical Best predictor based on a two-fold nested error model. The introduced statistical methodology is applicable to poverty indicators and more generally to separable nonlinear parameters paper. We use a parametric bootstrap method to estimate the mean squared error of the empirical best estimators. We also study small sample properties of these estimators by model-based simulation studies. The simulations show reductions in mean squared error relative to direct area-specific estimators and other estimators obtained by “simulated” censuses. We also apply the proposed method to estimate poverty incidences, poverty gaps, and the corresponding mean squared errors in Spanish provinces by gender. For the Spanish data, results show a significant reduction in coefficient of variation of the proposed empirical best estimators over direct estimators.
27 November 2015, 12:30, 1st floor Seminar room, IEMath-GR
