Recent Advances in Semiparametric Bayesian Function Estimation

Common nonparametric curve fitting methods such as spline
smoothing, local polynomial regression and basis function
approaches are now well developed and widely applied. More
recently, Bayesian function estimation has become a useful
supplementary or alternative tool for practical data analysis,
mainly due to breakthroughs in computerintensive inference via
Markov chain Monte Carlo simulation. This paper surveys recent
developments in semiparametric Bayesian inference for generalized
regression and outlines some directions in current research.