It is current in LODO meta-cohorts to try to suppress inter cohort technical (wetlab/collection) differences by approaches such as LEfSE or other statistical method (such as presented by Vadim Püller). If LODO approach is generalized to folds, this could be used to affect weight on features to discourage the usage of confounding feature across folds.
It is current in LODO meta-cohorts to try to suppress inter cohort technical (wetlab/collection) differences by approaches such as LEfSE or other statistical method (such as presented by Vadim Püller). If LODO approach is generalized to folds, this could be used to affect weight on features to discourage the usage of confounding feature across folds.