This work introduces how possibilistic regression can be used in the case of non symmetrical triangular membership functions, building a system of regressions, so that suitable restrictions for each particular problem can be incorporated. We apply this methodology to the problem of ecological inference, in particular to the estimation of the electoral transition matrix. An experimentation with several examples shows the benefits of the new approach.
© 2008-2024 Fundación Dialnet · Todos los derechos reservados