| _version_ |
1860799554427813888
|
| building |
INTELEK Repository
|
| collection |
Online Access
|
| collectionurl |
https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection407072
|
| date |
2018-12-22 12:54:33
|
| eventvenue |
Kuching, Sarawak, Malaysia
|
| format |
Restricted Document
|
| id |
6455
|
| institution |
UniSZA
|
| originalfilename |
1370-01-FH03-FESP-19-22912.pdf
|
| person |
C.Y. Chen
|
| recordtype |
oai_dc
|
| resourceurl |
https://intelek.unisza.edu.my/intelek/pages/view.php?ref=6455
|
| spelling |
6455 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=6455 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection407072 Restricted Document Conference Conference Paper application/pdf 1 1.6 Adobe Acrobat Pro DC 20 Paper Capture Plug-in C.Y. Chen 2018-12-22 12:54:33 1370-01-FH03-FESP-19-22912.pdf UniSZA Private Access Robust bootstrapping panel data Bootstrapping is a powerful tool for approximating the distribution of complicated statistics based on independent and identically distributed data. A natural way to bootstrap beta coefficients for fixed effect regression is by using residual-based bootstrap. However, the method heavily suffers the effects caused by high leverage points (HLPs). Random sampling with replacement in bootstrapping will introduce more outliers in the sub-samples of a contaminated data which then cause the bootstrap distribution to break down. We propose robustly weighted bootstrapping procedure that we called Boot RDF which incorporates the use of Robust Diagnostic-F to identify HLPs. Robust weights are then determined based on robust location of each data point from central data. In this way, lower weights are assigned to any outlying observation which in turn will lower down their chances of being included in the subsamples. The performance of Boot RDF are evaluated and compared to the existing fixed design, residual-based bootstrap via Monte Carlo simulation and numerical examples. The robust properties hugely increases the efficiency of the proposed Boot RDF; translated in the results of this study. International Quantitative Research and Applications Conference 2018 (IQRAC2018) Kuching, Sarawak, Malaysia
|
| spellingShingle |
Robust bootstrapping panel data
|
| summary |
Bootstrapping is a powerful tool for approximating the distribution of complicated statistics based on independent and identically distributed data. A natural way to bootstrap beta coefficients for fixed effect regression is by using residual-based bootstrap. However, the method heavily suffers the effects caused by high leverage points (HLPs). Random sampling with replacement in bootstrapping will introduce more outliers in the sub-samples of a contaminated data which then cause the bootstrap distribution to break down. We propose robustly weighted bootstrapping procedure that we called Boot RDF which incorporates the use of Robust Diagnostic-F to identify HLPs. Robust weights are then determined based on robust location of each data point from central data. In this way, lower weights are assigned to any outlying observation which in turn will lower down their chances of being included in the subsamples. The performance of Boot RDF are evaluated and compared to the existing fixed design, residual-based bootstrap via Monte Carlo simulation and numerical examples. The robust properties hugely increases the efficiency of the proposed Boot RDF; translated in the results of this study.
|
| title |
Robust bootstrapping panel data
|
| title_full |
Robust bootstrapping panel data
|
| title_fullStr |
Robust bootstrapping panel data
|
| title_full_unstemmed |
Robust bootstrapping panel data
|
| title_short |
Robust bootstrapping panel data
|
| title_sort |
robust bootstrapping panel data
|