Online Program Home
  My Program

Abstract Details

Activity Number: 35 - Special Session: Section on Nonparametric Statistics Student Paper Competition
Type: Contributed
Date/Time: Sunday, July 30, 2017 : 2:00 PM to 3:50 PM
Sponsor: Section on Nonparametric Statistics
Abstract #322652
Title: Bayesian Nonparametric Variable Selection for Optimal Treatment Decisions
Author(s): Rui Li* and Brian Reich and Howard Bondell
Companies: North Carolina State University and NCSU and NC State University
Keywords: Bayesian nonparametric ; Qualitative interaction ; Variable selection ; Personalized medicine

Personalized medicine relies on the ability to prescribe patient-specific treatments. Given the large amount of information collected from patients, it is crucial to identify key variables that impact the optimal treatment decision through variable selection. Variable selection research largely focused on selecting variables that are important for prediction, yet less attention has been paid to identify variables that are important for decision making. In this paper, we propose a Bayesian nonparametric variable selection method to identify important variables in treatment decision making. We use Gaussian process to model the treatment responses and adopt the regret function to evaluate the optimal treatment policies from different models. To select the best subset of variables, a backward selection procedure is used and the stopping criterion is based on hypothesis testing. The performance of our method is evaluated in simulation studies and an application with AIDS Clinical Trial Group Study.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

Copyright © American Statistical Association