Statistical Modelling 13 (4) (2013), 305–315

Discussion of "Beyond mean regression"

Peter J Green
University of Bristol and University of Technology,
Sydney,
Australia
e-mail: P.J.Green@bristol.ac.uk

Abstract:

Methodology for regression beyond the mean has been a goal of researchers for many years. This discussion provides some additional context for the important ideas in the present paper, by recounting some of the historical background to the GAMLSS approach and pointing to the power and appeal of fully probabilistic regression analysis in the setting of Bayesian nonparametrics.

Keywords:

back-fitting; Bayesian nonparametrics; Dirichlet process; LMS method; partial splines; penalised likelihood
back