From: A deep learning model incorporating part of speech and self-matching attention for named entity recognition of Chinese electronic medical records
P
R
F1
N-gram-Based RNN-CRF [23]
0.5254
0.4056
0.4578
Character-Based LSTM-CRF [11]
0.6568
0.6904
0.6732
Attention-Based CNN-LSTM-CRF [21]
0.7224
0.7248
0.7236
SM-LSTM-CRF
0.8054
0.7961
0.8007
