Actually we do not care here what data type and actual data is in the "outcome" column. It can be multilabel character strings integer, float, or anything, really. We just put this column aside and do not anonymize it. So maybe we should reflect this explicitly in the code and documentation.
Actually we do not care here what data type and actual data is in the "outcome" column. It can be multilabel character strings integer, float, or anything, really. We just put this column aside and do not anonymize it. So maybe we should reflect this explicitly in the code and documentation.