VARIANCE ESTIMATORS USING NON-PARAMETRIC APPROACH UNDER DIFFERENT RANKED SET SAMPLING SCHEMES

Abstract

Estimation of variance is a commonly discussed topic under simple random sampling (SRS) scheme. The current article deals the issue of variance estimation utilizing supplementary information with the nonparametric approach under different ranked set sampling (RSS) schemes. We propose a class of nonparametric variance estimators utilizing kernel regression [1] with different bandwidths (Plug-in and Cross-validation), under RSS schemes. Simulation study is provided utilizing diverse data sets. The comparison of simulation results has been made between the members of the proposed class with respect to the unbiased variance estimator.