Research Article
Multi-Self-Attention for Aspect Category Detection and Biomedical Multilabel Text Classification with BERT
Table 3
The test set results for SemEval-2014 task 4 Subtask 3: ACD TASK
| Model | Precision | Recall | F1 |
| ML-KNN | 70.77 | 56.06 | 62.09 | Rank SVM | 86.57 | 70.48 | 77.41 | TextCNN | 85.27 | 78.77 | 79.44 | BiLSTM-Attention | 81.20 | 80.67 | 79.28 | XRCE | 83.23 | 81.37 | 82.29 | Attention-XML | 84.73 | 82.24 | 83.43 | BERT base | 85.20 | 90.43 | 85.81 | BERT Att | 85.82 | 90.46 | 86.29 | BERT-MSA | 85.88 | 91.11 | 86.61 |
|
|