A Bayesian Method for Simultaneous Registration and Clustering of Functional Observations

• Activity Number: 365
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Zizhen Wu*
• Keywords: functional data ; curve registration ; time warping

Abstract

We develop a Bayesian method that simultaneously registers and clusters functional data of interest. Unlike other existing methods, which often assume a simple translation in the time domain, our method uses a discrete approximation generated from the family of Dirichlet distributions to allow warping functions of great flexibility. Under this Bayesian framework, a MCMC algorithm is proposed for posterior sampling. We demonstrate this method via simulation studies and applications to growth curve data and cell cycle regulated yeast genes.