Research Article
Multi-Instance Multilabel Learning with Weak-Label for Predicting Protein Function in Electricigens
Table 3
Performance of the MIMLwel methods with different weak-label ratios on two datasets.
| | Datasets | W.L.R. | HL↓ | maF1↑ | miF1↑ |
| | Geobacter sulfurreducens | 20% |
0.010 ± 0.002 | 0.003 ± 0.004 | 0.032 ± 0.035 | | 40% | 0.010 ± 0.002 | 0.009 ± 0.005 | 0.116 ± 0.038 | | 60% | 0.010 ± 0.002 | 0.016 ± 0.006 | 0.201 ± 0.034 | | 80% | 0.011 ± 0.001 | 0.019 ± 0.007 | 0.245 ± 0.050 |
| | Shewanella loihica PV-4 | 20% | 0.013 ± 0.002 | 0.009 ± 0.008 | 0.145 ± 0.111 | | 40% | 0.010 ± 0.002 | 0.005 ± 0.003 | 0.092 ± 0.039 | | 60% | 0.011 ± 0.003 | 0.010 ± 0.006 | 0.167 ± 0.072 | | 80% | 0.011 ± 0.003 | 0.011 ± 0.005 | 0.186 ± 0.043 |
|
|