2017-02-10

Jeremy Gaskins, PhD, Department of Bioinformatics and Biostatistics, University of Louisville

"Disease Classification from Longitudinal Data using Bayesian Nonparametrics"

Across many medical fields, developing an approach for disease classification is an important challenge. The usual procedure is to fit a model for the longitudinal response in the healthy population, fit a model for the longitudinal response in disease population, and apply Bayes’ Theorem to obtain disease probabilities given the responses. However, when substantial heterogeneity exists within each population, this Bayes classifier may have poor performance. In this work, we develop a new approach by fitting a Bayesian nonparameteric model for the joint outcome of disease status and longitudinal response. Due to the clustering induced by the Dirichlet process, our model will allow multiple subpopulations of healthy, diseased, and possibly mixed membership. We introduce an MCMC sampling scheme and discuss inference and prediction in our model. Our approach is demonstrated by predicting pregnancy outcome using hCG hormone levels in a sample of Chilean women being treated with assisted reproductive therapy.

Stay connected TwitterFacebookLinkedInYouTubeInstagram