In recent times, the role of statistical methods based on order statistics has become more and more significant in statistical inference. Let Y1 < Y2 < • • • < Yn be the order statistics corresponding to a random sample of a size n from a continuous distribution having probability density function f(x; e), e c S2. The purpose of this thesis is mainly to examine the procedures for estimating the parameter e using order statistics.The usual procedures for estimation of unknown parameters are based on the whole sample without taking into account the order in which the sample is taken or without arranging the observations in order of magnitude. Order statistics and estimations based on order statistics are becoming more popular due to their frequent use in nonparametric inferences and in robust procedures. Procedures based on order statistics are particularly useful when the examined data contain one or more extreme values or outliers.This thesis will provide useful insight to the problem of estimation using order statistics. Some works in this field will be studied, reviewed and updated. Estimation based on order statistics using full and censored samples for small and large data sets will be investigated with reference to continuous distributions, such as the normal distribution. In particular, estimation problems as well as hypothesis testing for location and scale parameters of some continuous distributions and estimation of quantiles based on order statistics will be examined.Ball State UniversityMuncie, IN 47306
Identifer | oai:union.ndltd.org:BSU/oai:cardinalscholar.bsu.edu:handle/182836 |
Date | 03 June 2011 |
Creators | Diebolt, Daniel T. |
Contributors | Ali, Mir M. |
Source Sets | Ball State University |
Detected Language | English |
Format | iii, 43 leaves ; 28 cm. |
Source | Virtual Press |
Page generated in 0.0014 seconds