(Learn more about this study)
Initial CDAS Request Approval
Nov 4, 2020
Heterogeneous Odds Ratio Estimation
The project develops new methodological tools to detect heterogeneity in patients' responses to treatment as measured by the odds ratio. In particular, heterogeneous treatment effect estimation are the first step towards personalized medicine/interventions. To the best of our knowledge, we are the first to consider the heterogeneous odds ratio estimation task, which is surprising given the popularity of odds ratio in medicine and social sciences and the profusion of research of HTE estimation.
We validate our methodological contributions by analyzing data from political science and revenue management. The PLCO data would constitute an interesting application to medicine to showcase the broad applicability and use of our approach.
- Develop a novel recursive partitioning procedure to estimate heterogeneous odds ratio, while most of the current literature considered difference in responses instead of ratios.
- Validate the methodology on different data sets from varied scientific fields: operations management, political science and medicine.