|
Activity Number:
|
193
- Section on Medical Devices and Diagnostics: Student Paper Competition
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 8, 2022 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Medical Devices and Diagnostics
|
|
Abstract #322649
|
|
|
Title:
|
Deep Spectral Q-Learning in Infinite Horizon with Applications in Mobile Health
|
|
Author(s):
|
Yuhe Gao* and Chengchun Shi and Rui Song
|
|
Companies:
|
North Carolina State University and London School of Economics and Political Science and North Carolina State University
|
|
Keywords:
|
Dynamic Treatment Regime;
Mobile Health;
Mixed Frequency Data;
Principal Component Analysis;
Reinforcement Learning;
Precision Medicine
|
|
Abstract:
|
Dynamic treatment regimes assign personalized treatments to subjects in a multi-stage decision process. In a mobile-health setting, such regimes will be developed based on collected patient information over a long period of study. The covariates of patients involved in decision making often include data collected at different frequencies. In this paper, we propose a deep spectral Q-learning algorithm, which integrates Principal Component Analysis (PCA) with deep Q-learning to handle the mixed frequency data. The estimated regime obtained by this method is proven to converge to the optimal regime. Simulation studies show the advantages of this proposed method and we also demonstrate this method on a diabetes dataset.
|
Authors who are presenting talks have a * after their name.
