Two well-known biased approaches, as an alternative to the ordinary least squares (OLS) estimator, to estimate the regression coefficients, namely, the ridge estimators and the Stein-type estimators, are considered. In this paper, a method of specifying the parameter c (the shrinkage factor) is proposed here and evaluated in terms of mean square error (MSE) by simulation techniques. Comparisons are made with James-Stein [6] proposed choice of the shrinkage factor c. Results of the simulation study indicate that with respect to MSE criteria, the suggested estimator outperforms both the OLS estimator and that of James and Stein in all the cases that have been investigated.

multicollinearity, OLS, ridge and Shrunken estimators, simulation.