肝癌患者在线提问的命名实体识别研究:一种基于迁移学习的方法 *
|
陈美杉,夏晨曦
|
Identifying Entities of Online Questions from Cancer Patients Based on Transfer Learning
|
Meishan Chen,Chenxi Xia
|
|
表5 实例迁移实验结果对比(%)
|
|
|
模型 | 评价 指标 | k=0 | k=1 | k=2 | k=3 | k=4 | k=5 | k=6 | KNN-BERT- BiLSTM-CRF | P | 92.91 | 93.54 | 94.89 | 95.74 | 95.40 | 94.73 | 94.60 | R | 95.36 | 95.74 | 96.51 | 96.75 | 96.24 | 96.30 | 95.68 | F | 94.12 | 94.63 | 95.69 | 96.10 | 95.82 | 95.51 | 95.14 | KNN-Word2Vec-BiLSTM-CRF | P | 85.98 | 88.73 | 90.45 | 91.48 | 91.65 | 91.03 | 90.77 | R | 86.55 | 89.57 | 91.30 | 92.48 | 92.62 | 92.05 | 91.90 | F | 86.26 | 89.15 | 90.87 | 91.98 | 92.13 | 91.54 | 91.33 |
|
|
|