Table 4 Results obtained for average-pooling CNN on the test dataset with negative filtering
From: Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction
Classes | P | R | F1 |
|---|---|---|---|
Advise | 66.99% | 63.35% | 65.12% |
Effect | 58.14% | 63.03% | 60.48% |
Int | 66.67% | 31.25% | 42.55% |
Mechanism | 61.90% | 47.99% | 54.06% |
Overall | 61.70% | 55.35% | 58.35% |