Dear @timbergquist and L3C Support, I am reaching out with a question regarding model robustness to missing data. Today during office hours, you mentioned that not all information available in the training data will be available in the test data (I apologize if I?m misquoting). Will the holdout test data include an equivalent covariate set to the censored training data? In other words, should our models be robust to individual missing values, or should they be robust to entire covariates (columns) being missing in the test data? Thanks so much for your time and help. Best, Zach

Created by Zachary Butzin-Dozier zbutzin
Hi Tim, Thank you so much for your response. We appreciate your help. Best, Zach
Hi @zbutzin , They should be robust to missing values. It's not guaranteed that every clinical concept in the training data will be available in the testing data. The columns in the OMOP tables will be the same. Thanks! Tim

Test Data Characteristics page is loading…