University of British Columbia
Multicollinearity, autocorrelation, and ridge regression
Abstract
dc:descriptionThe presence of multicollinearity can induce large variances in the ordinary Least-squares estimates of repression coefficients. It has been shown that ridge regression can reduce this adverse effect on estimation. The presence of serially correlated error terms can also cause serious estimation problems. Various two-stage methods, have been proposed to obtain good estimates of the regression coefficients in this case. Although the multicollinearity and autocorrelation problems have long been recognized in regression analysis, they are usually dealt with separately. This thesis explores the joint effects of these two conditions on the mean square error properties of the ordinary ridge estimator as well as the ordinary least-squares estimator. We show that ridge regression is doubly advantageous when multicollinearity is accompanied by autocorrelation in both,the errors and the principal components. We then derive a new ridge type estimator that is adjusted for autocorrelation. Finally, using simulation experiments with different degrees of multicollinearity and autocorrelation, we compare the mean square error properties of various estimators.
Degree
thesis:*- Name thesis:degree_name
- Master of Science in Business - MScB
- Level thesis:degree_level
- master's
- Discipline thesis:degree_discipline
- Business Administration
- Grantor dc:publisher
- University of British Columbia
- Year dc:date
- 1979
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Hsu, Jackie Jen-Chy
Rights
dc:rights- Statement dc:rights
-
- For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.
- Language dc:language
- eng
Identifiers
dc:identifier.*- Handle dc:identifier
- http://hdl.handle.net/2429/21865