Abstract

The Ordinary Least Square (LS) estimator remains Best Linear Unbiased Estimator (BLUE) when all the assumptions surrounding it stay intact, but at an iota of violation of the assumptions, it becomes inefficient and unstable. Multicollinearity is one of the causes of the violation of the assumptions. The ridge regression estimator given by (Hoerl & Kannard (1970)) is consistently attractive shrinkage method to reduce the effects of multicollinearity. In ridge estimation, the estimation of ridge parameter (k) is vital. However, there is still no way to compute its optimal value. In this article, we suggest a new method of finding the regression parameters which completely avoids the computation of “k”. New method compared to ridge estimators at different choices ridge parameters by simulation techniques in terms of Mean Square Error (MSE). Monte-Carlo simulation technique indicated that the proposed estimator performs better than LS estimators as well as Hoerl and Kennard’s ridge estimator at 149 different choice of ridge parameters reviewed in this article.

Author: A .V. Dorugade

Received on: December, 2023

Accepted on: February, 2025