Functional Data Analysis Using Bayesian Nonparametric Methods

• Activity Number: 52
• Type: Invited
• Date/Time: Sunday, July 31, 2011 : 4:00 PM to 5:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Bani Mallick*+
• Companies: Texas A & M University
• Keywords: Dirichlet process ; Clustering ; High dimensional data ; Gibbs sampling

Abstract

We propose a nonparametric Bayes model for clustering of functional data. The local basis function based methodology is aimed at the resolution of generic global and local features during clustering, and is suitable for clustering high-dimensional data. Based on the Dirichlet process (DP) (and other generalized processes), the nonparametric Bayes model extends the scope of traditional Bayes regression methods to functional clustering and allows the elicitation of prior belief about the regularity of the functions and the number of clusters by suitably mixing the DP. Posterior inference is carried out by Gibbs sampling with conjugate priors, which makes the computation straightforward. We use simulated as well as real data sets to illustrate the suitability of the approach over other alternatives.