Fig. 3From: A deep learning model incorporating part of speech and self-matching attention for named entity recognition of Chinese electronic medical recordsThe structure of SM-LSTM-CRF modelBack to article page