Table 8 Impact of the different dictionary sizes on method performance
From: Multi-task learning for Chinese clinical named entity recognition with external knowledge
Dataset | Dictionary size | Precision | Recall | F-measure |
|---|---|---|---|---|
CCKS2017 | 70% | 91.79 | 91.61 | 91.70 |
80% | 91.83 | 91.69 | 91.76 | |
90% | 91.87 | 91.73 | 91.80 | |
100% | 91.91 | 91.78 | 91.84 | |
FCCd | 70% | 87.19 | 86.65 | 86.92 |
80% | 87.23 | 86.71 | 86.97 | |
90% | 87.28 | 86.75 | 87.01 | |
100% | 87.32 | 86.79 | 87.05 |