Speaker: Jae-Kwang Kim, Professor, Department of Statistics, Iowa State University

Title: Introduction to Dynamic Treatment Regime

Abstract: Dynamic treatment regime (DTR) is an emerging area of research that provides a statistical framework for precision medicine and prescriptive analytics. Using the potential outcomes and the potential time-dependent covariates under different treatment options, one can use the DTR technique to make optimal decision-making for treatment options. Under some identification assumptions, the causal parameters can be estimated from the observational data, and the optimal treatment regime can be estimated by backward induction. Several estimation methods will be reviewed. Connections with reinforcement learning will also be presented.