A seminar by Associate Professor Shonosuke Sugusawa from University of Tokyo
Title: Bayesian causal synthesis for meta-inference on heterogeneous treatment effects
Abstract: Heterogeneous treatment effect (HTE) is widely known as a key measure in causal inference. There are now many techniques to estimate and inference on HTE based on randomized clinical trial or observational data. However, it is recognized that the performance of estimation methods can be different depending on the underlying data generating mechanism. Here we introduce a novel Bayesian framework called Bayesian causal synthesis (BCS) to combine multiple HTE estimators to make meta-inference on HTE. We provide a fast posterior computation algorithm and prove that the proposed BCS framework provides consistent estimator of HTE. We also demonstrate its numerical performance through simulation and empirical studies.
For further information, please contact RSFAS Seminars.
All information collected by the University is governed by the ANU Privacy Policy.