Research Article
Multifeature Named Entity Recognition in Information Security Based on Adversarial Learning
Table 2
Comprehensive performance evaluation for each model.
| | Method | Precision | Recall | F1 |
| | BiLSTM-Attention-CRF | 88.5% | 79.9% | 84.4% | | CRF-MA | 88.4% | 75.6% | 81.5% | | Dawid and Skene-LSTM | 79.3% | 75.1% | 77.1% | | BiLSTM-Attention-CRF-VT | 75.5% | 75.6% | 75.6% | | BiLSTM-Attention-CRF-crowd | 89.1% | 90.0% | 89.0% |
|
|
