Main Article Content
Adaptive Nonparametric Variance Estimation for a Ratio Estimator
Abstract
Kernel estimators for smooth curves require modifications when estimating near end points of the support, both for practical and asymptotic reasons. The construction of such boundary kernels as solutions of variational problem is a difficult exercise.
For estimating the error variance of a ratio estimator, we suggest an alternative estimation procedure using the theory of local linear regression. The proposed estimator adapts robustly to both interior and boundary points. We also derive the asymptotic mean square error of the new estimator and conditions under which it is efficient.
Journal of Agriculture, Science and Technology Vol.3(1) 2001: 34-42
For estimating the error variance of a ratio estimator, we suggest an alternative estimation procedure using the theory of local linear regression. The proposed estimator adapts robustly to both interior and boundary points. We also derive the asymptotic mean square error of the new estimator and conditions under which it is efficient.
Journal of Agriculture, Science and Technology Vol.3(1) 2001: 34-42