Diagnostic analysis of autism data using partially varying-coefficient mixed models under elliptical distributions

Abstract

In this degree work we extend the partially varying-coefficients mixed model with Normal er- rors to symmetric errors in order to permit distributions with heavier and lighter tails than the Normal distribution. This class of elliptical models includes all symmetric continuous distribu- tions, such as Student’s t, Pearson VII, Power Exponential and Logistic, among others. Estimation is performed by maximum penalized likelihood method and by using smoothing splines, also, a diagnostic analysis is performed by considering different perturbation schemes, in addition to a residual analysis. A real data set is analyzed using partially varying-coefficients mixed models with Normal errors and then under heavy-tailed distributions