A Nonparametric Bayesian Method for Hierarchical Clustering of Longitudinal Data

• Activity Number: 26
• Type: Contributed
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Yan Ren*+ and Siva Sivaganesan and Mekibib Altaye and Jane C. Khoury and Rhonda Vandyke
• Companies: University of Cincinnati and University of Cincinnati and Cincinnati Children's Hospital Medical Center and Cincinnati Children's Hospital Medical Center and Cincinnati Children's Hospital Medical Center
• Keywords: Bayesian ; Clustering ; Dirichlet process (DP) ; Gibbs sampler ; Longitudinal data ; Mixture model

Abstract

In longitudinal cohort studies, we are often interested in simultaneously clustering observations at patient- and time-levels. Current approaches, such as hierarchical and nested Dirichlet processes (DPs), are not applicable for this problem. Through the use of a specific base measure, we propose a more suitable method that improves upon the multivariate DP mixture model. A well-known MCMC algorithm, Gibbs sampler, is implemented for the Bayesian posterior distributions and estimates. The model is evaluated through simulation studies of multivariate data with different covariance specifications. Performance is assessed by using the correct classification rate and 95% credible intervals for parameter estimates. We also compare patient-level classification of our method to other approaches, including classification of intercepts and slopes based on linear regressions and multivariate DP mixture models. We illustrate the method with data from a prospective longitudinal study on maternal glycemic control, tracking the glucose levels and neonatal outcomes of 160 diabetic women during pregnancy.