Prediction in the Presence of Missing Covariates
share
In many applied fields incomplete covariate vectors are commonly encountered. It is well known that this can be problematic when making inference on model parameters, but its impact on prediction performance is less understood. We develop a method based on covariate dependent partition models that seamlessly handles missing covariates while avoiding completely any type of imputation. The method we develop is not only able to make in-sample predictions, but out-of-sample as well even if the missing pattern in the new subjects' incomplete covariate vector was not seen in the training data. Further, both categorical and continuous covariates are permitted. Our proposal fares well when compared to other alternatives based on imputations. We illustrate the method using simulation studies and an ozone dataset.
READ FULL TEXT