Table 5 Results for the artificial single serotype training data
From: A comparison of machine learning and Bayesian modelling for molecular serotyping
Dataset | BM | GBM | GBM Artf. Train Data | ||||
|---|---|---|---|---|---|---|---|
no filt. | Mixtures | 17 | (13.2-21.2) | 5.3 | (3.2-7.9) | 53 | (47.7-58.3) |
| Â | Singles | 8.4 | (7.5-9.3) | 9.8 | (8.8-10.8) | 39 | (37.4-40.6) |
| Â | Combined | 9.4 | (8.5-10.3) | 9.3 | (8.4-10.2) | 40.7 | (39.1-42.3) |
50% filt. | Mixtures | 14 | (10.5-17.9) | 4.4 | (2.5-6.8) | 45 | (39.7-50.3) |
| Â | Singles | 5.0 | (4.3-5.7) | 1.7 | (1.3-2.2) | 8.4 | (7.5-9.3) |
| Â | Combined | 6.1 | (5.4-6.9) | 2.0 | (1.6-2.5) | 12.8 | (11.8-13.9) |