Advances and Applications in Statistics
Volume 34, Issue 1, Pages 39 - 48
(May 2013)
|
|
THE HELPFUL RIDGE REGRESSION ESTIMATOR
Ghadban Khalaf
|
Abstract: In this paper, we consider the problem of estimating the regression coefficients in a multiple linear regression model in which the design matrix is singular. Under the assumption of normality, we propose a ridge regression parameter (k) and through simulation, the proposed estimator is proved to be uniformly better than the ordinary least squares (OLS) estimator, that is, minimax in term of standard deviation (S.D), and therefore, it is useful in the multicollinearity cases. |
Keywords and phrases: multicollinearity, OLSestimator, ridge regression, simulation. |
|
Number of Downloads: 358 | Number of Views: 1283 |
|