Improving the Performance of Least Squares Estimator in a Nonlinear Regression Model

摘要

In this study, we proposed novel preliminary test and shrinkage strategies for efficient estimation for multiple nonlinear regression models of the Cobb-Douglas type. A Monte Carlo simulation was conducted to evaluate the performance of the proposed estimators. To assess their practicality, they were applied to a real dataset. In simulations, the submodel estimator outperformed the other estimators when the sub-space information was true. The preliminary test and shrinkage strategies outperformed the widely-used Least Absolute Shrinkage and Selection Operator in this case. Conversely, the shrinkage estimators outperformed the other estimators in most of the parameter space when the subspace information was untrue.