Fig. 5From: A deep learning model incorporating part of speech and self-matching attention for named entity recognition of Chinese electronic medical recordsPerformance comparison of different input methods on the five types of entitiesBack to article page