Yanyuan Ma, PhD, Professor of Statistics, Penn State University
In studies ranging from clinical medicine to policy research, complete data are usually available from a population which only has partial data. We consider the setting under the label shift assumption; i.e., the conditional distribution of covariates given response is the same in the two populations. We propose an estimation procedure that only needs some standard nonparametric technique to approximate the conditional expectations and is doubly flexible, a procedure that is more flexible than the well-known doubly robust estimation. We develop the large sample theory for the proposed estimator, and examine its finite-sample performance through simulation studies as well as an application to the MIMIC-III.