|
Activity Number:
|
580
- Statistical and Computational Challenges in Nonparametric Learning
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Thursday, August 6, 2020 : 3:00 PM to 4:50 PM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
|
Abstract #309623
|
|
|
Title:
|
Learning Non-Monotone Optimal Individualized Treatment Regimes
|
|
Author(s):
|
Trinetri Ghosh* and Yanyuan Ma and Wensheng Zhu
|
|
Companies:
|
Pennsylvania State University and The Pennsylvania State University and Northeast Normal University
|
|
Keywords:
|
Cross-validation;
Double-robust;
Optimal individualized treatment regimes;
Propensity score;
Value function
|
|
Abstract:
|
We propose a new method to estimate the optimal treatment regime under minimal conditions when two different treatments can be selected. We consider a robust estimating equation approach to protect against misspecification of the propensity score function, the outcome regression model for the non-treated group and the non-monotonic treatment difference model. Nonparametric smoothing is incorporated to estimate the treatment difference model. We then identify the optimal treatment accordingly to maximize the value function. Theoretical properties of the treatment assignment strategy are established and we illustrate the performance and effectiveness of our proposed estimators through simulation and data applications.
|
Authors who are presenting talks have a * after their name.
