Resampling Techniques

Abstract

Resampling techniques are PC escalated procedures which include drawing multiple new samples called resamples from the data sample at hand. To investigate and estimate various properties of any original estimator used for estimating a population parameter of interest, all estimators of interest are determined from each of the resamples. These resampling based estimators are then analyzed to estimate the properties of the original estimator. Utilization of resampling methods is justified in circumstances where assurance of properties of estimators is not straight for example in cases where there are violations of distributional assumptions, small sample sizes etc. Moreover, the classical approach has been mostly developed for random samples; however, for nonrandom samples, resampling techniques are sometimes more suitable.