Longitudinal Joint Clustering via Dirichlet Process Mixture Model

• Activity Number: 404 - Bayesian Clustering and Classification
• Type: Contributed
• Date/Time: Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Hongmei Zhang* and Shengtong Han and Hasan Arshad
• Companies: University of Memphis and University of Chicago and University of Southampton
• Keywords: Longitudinal clustering ; Dirichlet Process ; Bayesian methods ; Joint clustering

Abstract

This piece of work proposes a clustering approach based on Dirichlet process (DP) mixture model. To model both time invariant and temporal patterns, different from other existing clustering methods, the proposed semi-parametric model has the ability to differentiate common and unique longitudinal patterns among observations, and jointly cluster subjects and variables. Via joint clustering, the intrinsic complex shared patterns among subjects and among variables are captured. The number of joint clusters and cluster assignments are self-controlled with the use of DP. Simulation studies illustrate the effectiveness of the proposed method. An application to wheal size allergic sensitization data is used to demonstrate the method.