Analysis of Simultaneous Equation Model (SEM) on Non normally Response used the Method of Reduce Rank Vector Generalized Linear Models (RR-VGLM)

Abstract

Multivariate linear regression is a statistical analysis methods used to data in the case of multiple response variables associated with several predictor variables. In this method of analysis there is an assumption of matrix coefficient regression must be full rank. In the case of simultaneous equations, full rank condition is not fulfilled. Consequently, to analyze that case is not possible because it would produce a regression coefficient that is very large so needed reduction in rank. Reduce Rank Regression (RRR) Method is an alternative method in the case where this method if there is a weak regression coefficients will be cut. However, Reduced rank regression method only applies in response which continuous and normal in econometric data analysis and others. Therefore, to overcome that problem so introduced to f analysis method of Reduce Rank Vector Generalized Linear Model (RR-VGLM). This article will discuss simultaneous equations with non-normal variable response using RR-VGLM by simulating non normal conditions.