Conferência Prêmio IASI

Scale Mixtures of Skew-Normal Distribution with Applications in Regression Models

Victor H. Lachos (UNICAMP)

Scale mixtures of skew--normal distribution is a class of asymmetric distributions which includes the skew-normal one as a special case. In this paper, I explore the use of the EM-algorithm to develop likelihood analysis of two multivariate models of practical interest. Moreover, full maximum likelihood inference is adopted since some simplifying and unifying results are noted with the observed information matrix, which is derived analytically. The procedures are discussed with emphasis on the skew--t, on the skew--slash, and on the skew—contaminated normal distributions. Results obtained from two real data sets are reported illustrating the usefulness of the proposed methodology.