Table 7 The confusion matrix of model-term on the test set with basic, training set optimisation fine-tuning strategy
From: Automatic consistency assurance for literature-based gene ontology annotation
| Â | Predicted labels | |||
|---|---|---|---|---|
Consistent | (A) | (B) | (C) | |
True Labels | ||||
Basic system | ||||
 Consistent | 850 | 56 | 656 | 17 |
 (A) Contradictory | 15 | 61 | 52 | 0 |
 (B) Over-specific | 186 | 52 | 972 | 21 |
 (C) Unsupportive | 156 | 42 | 354 | 1027 |
+Training set optimisation | ||||
 Consistent | 1164 | 56 | 309 | 17 |
 (A) Contradictory | 47 | 69 | 12 | 0 |
 (B) Over-specific | 239 | 39 | 933 | 20 |
 (C) Unsupportive | 199 | 44 | 379 | 957 |