Manuscript Number : IJSRSET1849101
A Comparison of Cluster Method and Nearest Neighbor Method for Non-sample Area in the Small Area Estimation
Authors(3) :-Annastasia Nika Susanti, Kusman Sadik, Anang Kurnia
Small area estimation is an indirect method to estimate the parameter of a population by using the model approach. The problem that often arises in the small area estimation is non-sampled area then the area random effect of non-sampled areas is can not estimate because no sample units are available in these areas. This paper proposed a method to solve the non-sampled area problem by adding the cluster information and by using the nearest neighbor area to estimate the area random effect through the Fast Hierarchical Bayes (FHB) approach. These methods are compared by using the simulation study and the evaluation is based on the Absolute Relative Bias (ARB) and Relative Root Mean Square Error (RRMSE). The result shows that the estimation by using the cluster method has smaller ARB and RRMSE values than the estimation by using the nearest neighbor area in various sample sizes and various population sizes. Then it can be said that cluster method is better to provide the estimators of non-sampled area than the nearest neighbor area method.
Annastasia Nika Susanti
Small Area Estimation, Non-Sampled Area, Cluster Method, The Nearest Neighbor Area, Poverty Indicators.
Publication Details
Published in :
Volume 4 | Issue 9 | July-August 2018 Article Preview
Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia
Kusman Sadik
Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia
Anang Kurnia
Department of Statistics, Bogor Agricultural University, Bogor, West Java, Indonesia
Date of Publication :
2018-08-30
License: This work is licensed under a Creative Commons Attribution 4.0 International License.
Page(s) :
463-468
Manuscript Number :
IJSRSET1849101
Publisher : Technoscience Academy
Journal URL :
https://ijsrset.com/IJSRSET1849101