Between Explorative and Confirmative Estimation Methods for the Structural Equation Models
Abstract
En
The aim of this work is to show an integrated approach for using the Structural Equation Model (SEM) with an information theoretic proposal when the classic approach doesn’t work. The estimation methods used are the Partial Least Squares (PLS), for a non-parametric and explorative analysis and the Maximum Likelihood Estimation (MLE) method, for parametric and confirmative analysis. In this respect, the innovatory aspects are the follows: (i) the estimation methods are used as an statistical integrated approach, in order to apply the potentiality and the main characteristics of both of them; (ii) the extension of the SEM model in case both of the estimation methods cannot perform well the data, by considering the information theory of the Generalized Maximum Entropy (GME).
The aim of this work is to show an integrated approach for using the Structural Equation Model (SEM) with an information theoretic proposal when the classic approach doesn’t work. The estimation methods used are the Partial Least Squares (PLS), for a non-parametric and explorative analysis and the Maximum Likelihood Estimation (MLE) method, for parametric and confirmative analysis. In this respect, the innovatory aspects are the follows: (i) the estimation methods are used as an statistical integrated approach, in order to apply the potentiality and the main characteristics of both of them; (ii) the extension of the SEM model in case both of the estimation methods cannot perform well the data, by considering the information theory of the Generalized Maximum Entropy (GME).
DOI Code:
�
Keywords:
Structural Equation Model; Partial Least Squares; Maximum Likelihood Estimation; Generalized Maximum Entropy
Full Text: PDF
