A simulation study of biased estimators against the ordinary least squares estimator
Titel:
A simulation study of biased estimators against the ordinary least squares estimator
Auteur:
Thiart, C. Dunne, T.T Troskie, C.G Chalton, D.O
Verschenen in:
Communications in statistics
Paginering:
Jaargang 22 (1993) nr. 2 pagina's 569-589
Jaar:
1993
Inhoud:
A simulation study was performed to examine the relative efficiency of thirteen estimators against OLSE across thirty combinations of orientation, variance and known parameter vector, encompassing several different degrees of collinearity. Estimation procedures include principal components, ridge and generalized ridge, jackknife, shrinkage and fractional principalcomponent methods. Overall, in the face of collinearity the biased estimators perform better than OLSE, but no 'optimal' estimator emerges, and a suitable choice of estimator will depend on the circumstances