Table 3. Results about redundancies when applied these methods on synthetic dataset S2 (Mixture dataset) and S3 (Gamma dataset) to select 10 features.
In columns for “LEfSe,” “Metastats,” “Wilcoxon” and “MetaBoot,” the values were the non-redundancy rate (Eq. (2)) of non-redundant biomarkers with standard deviation of 1.
| Dataset | LEfSe | Metastats | Wilcoxon | MetaBoot |
|---|---|---|---|---|
| S2 (Mixture dataset) | 36.0 ± 5.5 | 26.0 ± 5.5 | 38.0 ± 8.4 | 42.0 ± 4.5 |
| S3 (Gamma dataset) | 46.0 ± 11.4 | 31.4 ± 9.0 | 50.0 ± 12.2 | 50.9 ± 8.1 |