Classification accuracy in terms of average K-category correlation coefficient (KCCC) using weighted and unweighted PhyloRelief, LEfSe using OTUs and classified taxa, RF and MetaPhyl.

For PhyloRelief, three value of k (k = 2,3,4) are shown. When feature selection was performed using PhyloRelief, LEfSe and RF, the RF classifier was used. For each of these algorithms we report the cross-validation accuracy in terms of average KCCC, the Standard Error and the number of features selected in the final model using the complete dataset (in parentheses). For PhyloRelief and RF the number of features was selected by a nested cross validation loop. For each dataset, the maximum KCCC value is marked in bold.
Classification accuracy in terms of average K-category correlation coefficient (KCCC) using weighted and unweighted PhyloRelief, LEfSe using OTUs and classified taxa, RF and MetaPhyl.