Some Improved Generalized Ridge Estimators and their comparison

Lukman, A. F. and Haadi, A. and Ayinde, K. and Onate, C.A and Gbadamosi, Babatunde and Oladejo, N. K. (2018) Some Improved Generalized Ridge Estimators and their comparison. WSEAS TRANSACTIONS on MATHEMATICS, 17. pp. 369-376. ISSN 2224-2880

[img] Text
a875906-1032.pdf - Published Version

Download (403kB)


The problem of multicollinearity is often encountered in time series data since explanatory variables included in the model often share a common trend. Various methods exist in literatures to handle this problem. Among them is the most widely used ridge regression estimator which depends on the ridge parameter. This estimator can be subdivided into either generalized ridge or ordinary ridge estimators. Variance inflation factor is introduced to replace eigenvalue in the generalized ridge estimator proposed by Lawless and Wang (1976). Through this modification some new generalized ridge parameters are proposed and investigated via simulation study. The performances of these proposed estimators are compared with the existing ones using mean square error. Results show that the proposed estimators perform better than the existing ones. It is evident that increasing the level of multicollinearity and number of regressors has positive effect on the MSE. Also, the performance of the estimators depends on the level of error variances

Item Type: Article
Subjects: Q Science > QA Mathematics
Date Deposited: 07 May 2019 12:15
Last Modified: 13 Sep 2019 11:11

Actions (login required)

View Item View Item