A COMPARISON OF DIFFERENT CRITERIA TO CONSTRUCT REGRESSION MODEL EMPLOYING THE BOX-COX AND COLE GREEN TRANSFORMATION
This article introduces an algorithm that utilizes power transformation to estimate a nonlinear regression model for Cole Green and Box-Cox transformation. The algorithm outlines steps for selecting an optimal powers parameter estimate, employing the Akaike Information Criterion and Bayesian Information Criterion, statistical modeling efficiency criteria are incorporated to complement the traditional method. Additionally, Decision rules include the adjusted coefficient of determination Maximum Likelihood Estimator and the F-statistics test. The proposed algorithm is applied to real data, and the conclusion emphasizes the feasibility of obtaining various options exist for selecting the optimal power parameter. However, attaining a singular optimal value that meets both estimation and decision criteria methods is deemed impractical.
Cole Green Transformation, Box-Cox Transformation, Adjusted R-Square, and Akaike Information Criterion and Bayesian Information Criterion.