Analysis of Censored Sample Population with GA-SVM

Abstract

This paper is intended to propose a class of shrunken estimators for kth power of scale parameter in censored samples from one-parameter exponential population when some apriori or guessed value of the parameter is available besides the sample information and analyses their properties. The proposed class of Shrunken estimator is compared with usual unbiased estimator and minimum mean square error (MMSE) estimator. Eventually, empirical study is carried out to exhibit the performance of some Shrunken estimators of the proposed class over the MSME estimator. It is found that certain of these estimators substantially improve the classical estimators even for the guessed values of the kth power of scale parameter much away from the true value, specially for censored samples with small sizes.