Nonparametric Bayes kernel-based priors for functional data analysis

Richard F. MacLehose and David B. Dunson

Biostatistics Branch, NIEHS

September, 2007

We focus on developing nonparametric Bayes methods for collections of dependent random functions, allowing individuals curves to vary flexibly while adaptively borrowing information. A prior is proposed, which is expressed as a hierarchical mixture of weighted kernels placed at unknown locations. The induced prior for any individual function is shown to fall within a reproducing kernel Hilbert space. We allow flexible borrowing of information through the use of a hierarchical Dirichlet process prior for the random locations, along with a functional Dirichlet process for the weights. Theoretical properties are considered and an efficient MCMC algorithm is developed, relying on stick-breaking truncations. The methods are illustrated using simulation examples and an application to reproductive hormone data.

Keywords: Dirichlet process; Functional data analysis; Kernel smoothing; Mixture model; Random curve; RKHS.


The manuscript is available in PDF formats.