@meganomaly can we use test set missingness patterns/ statistics to derive training model inputs?
cool
cool