Mislabeled metadata are especially problematic in large-scale collaborations where data analysis is far removed from data generation: researchers cannot reconfirm sample labels or resequence ...

Error-Correcting Output Codes (ECOC) provide a robust framework for decomposing multi-class classification challenges into multiple binary sub-problems. By ...

Measurement error is often a severe problem in longitudinal data such as labor-market flows. Two correction methods have been used for labor-market flows, but the ...