SIMPLIFICACIÓN DEL GRADIENTE DE LA FUNCIÓN DE MÁXIMA VEROSIMILITUD DEL ANÁLISIS FACTORIAL CONFIRMATORIO

This paper present the process of simplification of the function maximum likelihood gradient, used in the estimation of confirmatory factor analysis. The gradient obtained is presented infunction of CFA tradition al matrix: Λ, Φ y Θe (regression coefficients, variances of latent variables and error variances). This simplification was performed using matrices derivation laws and yielded an expression for the gradient easy for programing.