|
Activity Number:
|
475
- Statistical Computing and Inference
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 2, 2017 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
|
Abstract #323303
|
|
|
Title:
|
SUFFICIENT WEIGHTED BOOTSTRAPPING
|
|
Author(s):
|
Khairul Islam* and Sarjinder Singh
|
|
Companies:
|
Eastern Michigan University and Texas A&M University-Kingsville
|
|
Keywords:
|
Bootstrapping ;
estimation of mean ;
auxiliary variable ;
sufficient weighted bootstrapping ;
simulation
|
|
Abstract:
|
Singh and Sedory [1] have introduced the idea of sufficient bootstrapping, which is based on retaining only distinct individual responses from a simple random and with replacement sampling (SRSWR). The idea of usage of the distinct units in SRSWR at the estimation stage is well known from the work of Raj and Khamis [2]. Basu [3] has investigated the theory of distinct units under probability proportional to size and with replacement sampling (PPSWR). Singh [4] has studied a weighted bootstrapping in the presence of auxiliary information and named as saddlestrapping because there exists a saddlepoint based on the correlation between the study and auxiliary variable. In this paper, we extend the idea of distinct units in PPSWR sampling to construct a new Sufficient Weighted Bootstrapping.
|
Authors who are presenting talks have a * after their name.
