JSM 2012 Home

JSM 2012 Online Program

The views expressed here are those of the individual authors and not necessarily those of the JSM sponsors, their officers, or their staff.

Online Program Home

Abstract Details

Activity Number: 105
Type: Invited
Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Mining
Abstract - #303585
Title: New Reinforcement Learning Methodology for Personalized Medicine
Author(s): Michael Kosorok*+
Companies: The University of North Carolina at Chapel Hill
Address: 3101 McGavran-Greenberg Hall, CB 7420, Chapel Hill, NC, 27599, United States
Keywords: dynamic treatment regimes ; personalized medicine ; reinforcement learning ; machine learning ; Q-learning ; A-learning

Finding individualized dynamic treatment regimes is a fundamentally important challenge in personalized medicine research. Reinforcement learning is a primary statistical learning tool used in this setting, with Q-learning and A-learning being the most common forms of reinforcement learning. We propose an alternative form of reinforcement learning which yields dramatically improved performance for high dimensional training data arising from randomized clinical studies. We prove consistency, derive rates of convergence, and compare with alternative approaches using several revealing numerical studies.

The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2012 program

2012 JSM Online Program Home

For information, contact jsm@amstat.org or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.