Abstract:
|
Calibration estimation, a technique of adjusting the sampling weights to match the unknown population totals of auxiliary variables, is a popular method of estimation in survey sampling. When the auxiliary vairbales are observed for all units in the finite population, one can apply the model calibration of Wu and Sitter (2001) using the working outcome model. In this paper, we develope a kernel-based nonparametric calibration method that does not require an explicit outcome model. The proposed method is a function calibration employing infinite-dimensional reproducing kernel Hilbert space (RKHS). Numerical algorithms are developed and implemented to solve the optimization problem in the function calibration, and some asymptotic results are presented as well. Furthermore, under the nonparametric working model, the proposed calibration estimator attains the Godambe-Joshi lower bound asymptotically. Simulation results are presented to compare the proposed method with other calibration methods.
|