Multimodal prediction of conversion to Alzheimer's disease based on incomplete biomarkers^∗

Abstract

Extensive multimodal and incomplete data can be adequately handled by a combination of missing data substitution, feature selection, and classification.