Table 4 Performance of different models
From: Child-Sum EATree-LSTMs: enhanced attentive Child-Sum Tree-LSTMs for biomedical event extraction
Dataset | Method | Precision (%) | Recall (%) | F1-score (%) |
|---|---|---|---|---|
MLEE | RecurCRF [29] | 81.12 | 79.15 | 80.28 |
Yan Wang [31] | 82.20 | 78.25 | 80.18 | |
Xinyu He [30] | 82.01 | 78.02 | 79.96 | |
Attentive Child_Sum Tree-LSTM | 82.95 (82.75 ± 0.19) | 80.62 (80.41 ± 0.21) | 81.77 (81.51 \(\pm \hspace{0.17em}\)0.19) | |
Child_Sum EATree-LSTM | 83.24 (83.00 ± 0.19) | 80.90 (80.71 ± 0.21) | 82.05 (81.96 \(\pm \hspace{0.17em}\)0.19) | |
BioNLP’09 | RecurCRF | 76.42 | 70.45 | 73.24 |
Attentive Child_Sum Tree-LSTM | 75.95 (75.71 ± 0.19) | 72.23 (72.01 ± 0.21) | 74.11 (73.90 \(\pm \hspace{0.17em}\)0.19) | |
Child_Sum EATree-LSTM | 76.84 (76.64 ± 0.19) | 73.35 (73.11 ± 0.21) | 75.05 (74.86 \(\pm \hspace{0.17em}\)0.19) |