Table 4.
Mean normalised classification rates over 4 multi-phenotype datasets according to performance
| Multiclass | DIGS | SG | Mean [ 28 ] | Median [ 28 ] | Per_pathway [ 72 ] | PCA [ 45 ] |
|---|---|---|---|---|---|---|
| 5-NN | 1 | 0.9532 | 0.8126 | 0.8090 | 0.8158 | 0.8696 |
| NN | 1 | 0.9488 | 0.9402 | 0.9334 | 0.8322 | 0.9585 |
| SMO | 1 | 0.9335 | 0.9372 | 0.9246 | 0.8521 | 0.9452 |
| HB | 1 | 0.9241 | 0.7518 | 0.7639 | 0.7893 | 0.8043 |
| Logistic | 1 | 0.8290 | 0.5614 | 0.5440 | 0.5589 | 0.6450 |
| Mean | 1 | 0.91772 | 0.80064 | 0.79498 | 0.76966 | 0.84452 |
The highest classification rate achieved across all competing methods is highlighted as bold for each classifier.