A test statistic for selection of multivariate linear regression models
Issued Date
2015
Issued Date (B.E.)
2558
Available Date
Copyright Date
Resource Type
Series
Edition
Language
eng
File Type
application/pdf
No. of Pages/File Size
185 leaves
ISBN
ISSN
eISSN
Other identifier(s)
b191876
Identifier(s)
Access Rights
Access Status
Rights
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Rights Holder(s)
Physical Location
National Institute of Development Administration. Library and Information Center
Bibliographic Citation
Citation
Srisuda Boonyim (2015). A test statistic for selection of multivariate linear regression models. Retrieved from: https://repository.nida.ac.th/handle/662723737/5758.
Title
A test statistic for selection of multivariate linear regression models
Alternative Title(s)
Author(s)
Advisor(s)
Editor(s)
item.page.dc.contrubutor.advisor
Advisor's email
Contributor(s)
Contributor(s)
Abstract
In this study, a test statistic used to select a multivariate linear regression model based on Mallows’s Cp with the same rationale as the SCp criterion from the
system of equations Vichit Lorchirachoonkul and Jirawan Jitthavech (2012: 2386- 2394) proposed. The D statistic, which is the difference between the modified Cp
statistics in the reduced model and in the full model, approximates to a standard normal distribution.
The modified C p statistic without the hypothesis testing, MC , and the proposed test statistic TD based on the percentage of selecting the model correctly were compared via a simulation study. Variable selection was carried out using backward elimination with five datasets consisting of 100 samples of size 200 and significance levels of 0.05 and 0.10, and the correlation between the equations was set at 0.3, 0.4, 0.5, 0.7, and 0.8, respectively. The multivariate linear regression full models consisted of two dependent variables, two relevant independent variables, and two irrelevant independent variables. In addition, the random disturbances were uncorrelated across observations in the same equation but contemporaneously correlated across equations. The results of the simulation study showed that the test statistic TD was able to select the model more often than the modified Cp criterion in all datasets, and, for both criteria, no under-fit models were selected.
The modified C p statistic without the hypothesis testing, MC , and the proposed test statistic TD based on the percentage of selecting the model correctly were compared via a simulation study. Variable selection was carried out using backward elimination with five datasets consisting of 100 samples of size 200 and significance levels of 0.05 and 0.10, and the correlation between the equations was set at 0.3, 0.4, 0.5, 0.7, and 0.8, respectively. The multivariate linear regression full models consisted of two dependent variables, two relevant independent variables, and two irrelevant independent variables. In addition, the random disturbances were uncorrelated across observations in the same equation but contemporaneously correlated across equations. The results of the simulation study showed that the test statistic TD was able to select the model more often than the modified Cp criterion in all datasets, and, for both criteria, no under-fit models were selected.
Table of contents
Description
Thesis (Ph.D. (Statistics))--National Institute of Development Administration, 2015