Estimation and prediction in the presence of an outlier under Type-II censoring |
| |
Authors: | Bahare Khatib Mostafa Razmkhah |
| |
Affiliation: | 1. Department of Mathematics and Statistics, University of Neyshabur, Neyshabur, Iran;2. Department of Statistics, Ferdowsi University of Mashhad, Mashhad, Iran |
| |
Abstract: | A single-outlier data set containing some independent random variables is considered such that all of observations expect one have the same distribution. To describe the model of interested, a location-scale family of distributions is used and the estimation problem of the parameters is studied when the data are collected under Type-II censoring scheme. Moreover, three different predictors are presented to predict the censored order statistics. They are also compared regarding both of mean squared prediction error and Pitman's measure of closeness criteria. The role of outlier parameter as well as censorship rate is studied on performance of proposed estimator and predictors. The results of the paper are illustrated via a real data set. Finally, some conclusions are stated. |
| |
Keywords: | Best linear unbiased estimator best linear unbiased (invariant) predictor censored data location-scale family Pitman's measure of closeness |
|
|