←Back
506 – New Ideas in Inference
Two Parameter Estimators: Biased and Almost Unbiased Estimation for Nonorthogonal Problems
Sponsor: Section on Statistical Computing
Keywords: linear regression model, multicollinearity, ridge regression, two-parameter estimator, almost unbiased estimator, chemical structure data sets
Muhammad Qasim
Jönköping University
In this paper, we consider the estimation of the parameter (β) in a classical linear regression model by combining the ridge and Liu estimators. The biased and almost unbiased two-parameter estimators are proposed. The necessary and sufficient conditions for the superiority of the proposed estimators over the existing estimators in terms of matrix mean squared error are derived. Besides, we suggest the algorithm for choosing the shrinkage parameters (k & d) for newly developed estimators. The performance of the estimators is gauged through Monte Carlo simulation and empirical application.