Distribuições combinadas

Abstract

It is proposed a possible new line of research in which distributions are combined. Combined distributions are an alternative for adjusting continuous data that present different probabilistic behaviors. A combined distribution called Gucumatz is developed, composed of the symmetrical Normal and Cauchy distributions. Some properties of this distribution are studied, maximum likelihood estimation of augmented data is performed (EM Algorithm), regression model and simulation studies are developed. A set of real data is adjusted using the Gucumatz distribution.