UNA IMPLEMENTACION DEL AJUSTE DE DATOS MEDIANTE MINIMOS CUADRADOS GENERALIZADOS NO LINEALES PARA FUNCIONES VECTORIALES

Abstract

n this communication, a critical survey mainly on generalized least squares (GLS) software is made, and afterwards, to help to overcome the software limitations (according to our objectives) an algorithm for vector valued data fitting by GLS is presented. Moreover, it is shown another algorithm to estimate the variance-covariance matrix that corresponds to vectorized observation matrix, with different options for users. Both algorithms were implemented on MATLAB Version 7.3.