Abstract:
|
In this article, an attempt was made to select a model for the 2015 and 2016 shrimp data in the Gulf of Mexico by considering several possibilities for the model. A general linear model (GLM) with three different methods for estimating the parameters were studied. The estimation methods considered were ordinary least square (OLS), generalized least square (GLS), and feasible generalized least square (FGLS). In the case of GLS, two different weights were selected for the improvement of the heteroscedasticity and the proper weight (s) was deployed. The third weight was selected through the application of FGLS. Analyses showed that only two of the three weights were effective in reducing the severity of the heteroscedasticity. Data sets were divided into training, validation, and testing. Stepwise, backward and forward methods with some statistics like ASE with proper hierarchies were deployed. Furthermore, the response variable in both data files was checked for the normality requirement using the Box-Cox transformation method. Analysis showed that the logarithmic transformation solved the normality in a satisfactorily. A few versions of the GLM were selected as possible candidates.
|