Table 3 List of the methods, stringency (high and low) and evaluation criteria used for both regression and classification
From: A random forest based biomarker discovery and power analysis framework for diagnostics research
RF methods | Stringency | Criteria | Regression | Classification | ||||
|---|---|---|---|---|---|---|---|---|
Simulation | Metabolomics | Lipidomics | Simulation | Lipidomics | Transcriptomics-1 | |||
RFE | High | TP/Known | 20 | 1 | 1 | 10 | 0 | – |
| Â | Â | FP/Novel | 0 | 5 | 0 | 0 | 0 | 0 |
|  | Low | TP/Known | 29 | 3 | 3 | 29 | 2 | – |
| Â | Â | FP/Novel | 19 | 91 | 8 | 201 | 18 | 14 |
Boruta | High | TP/Known | 20 | 3 | 2 | 11 | 2 | – |
| Â | Â | FP/Novel | 0 | 43 | 6 | 0 | 24 | 19 |
|  | Low | TP/Known | 29 | 3 | 3 | 29 | 3 | – |
| Â | Â | FP/Novel | 1 | 83 | 34 | 9 | 10 | 39 |
Permutation (Raw) | High | TP/Known | 29 | 2 | 2 | 19 | 2 | – |
| Â | Â | FP/Novel | 0 | 24 | 7 | 0 | 10 | 0 |
|  | Low | TP/Known | 29 | 3 | 3 | 29 | 3 | – |
| Â | Â | FP/Novel | 98 | 68 | 47 | 465 | 35 | 132 |
Permutation (Corrected) | High | TP/Known | 21 | 2 | 1 | 11 | 2 | – |
| Â | Â | FP/Novel | 0 | 1 | 0 | 0 | 6 | 0 |
|  | Low | TP/Known | 29 | 3 | 3 | 29 | 3 | – |
| Â | Â | FP/Novel | 8 | 48 | 26 | 110 | 46 | 66 |