Table 3 Test results of the two neural network models on given dataset
From: A hybrid approach for named entity recognition in Chinese electronic medical record
Entity name | BiLSTM-CRF | Attention-BiLSTM-CRF | ||||
|---|---|---|---|---|---|---|
Strict index (%) | Strict index (%) | |||||
P | R | F1 | P | R | F1 | |
Anatomy | 85.57 | 85.61 | 85.59 | 86.29 | 86.24 | 86.27 |
Surgery | 85.81 | 85.58 | 85.69 | 86.19 | 84.90 | 85.54 |
Drug | 94.92 | 78.11 | 85.70 | 89.73 | 85.98 | 87.81 |
Independent symptom | 92.45 | 89.52 | 90.96 | 91.93 | 90.20 | 91.05 |
Symptom description | 91.81 | 87,91 | 89.82 | 91.58 | 87.69 | 89.59 |
Total | 87.66 | 85.72 | 86.68 | 87.75 | 86.77 | 87.26 |