Named Entity Recognition of TCM Text Based on Bert Model

Aiming at the problem of fuzzy entity recognition and less labeled data in the field of traditional Chinese medicine, a named entity recognition model based on Bert-BiLSTM-CRF is constructed and tested on the corresponding data set. According to the text information, it is divided into five types of entities: symptoms, disease names, time, prescription names, and drug names. The results show that the model has the highest accuracy in identifying drug names. In order to further prove the superiority of this model, three groups of control groups composed of other models are set up. The model far exceeds the control group in terms of accuracy, recall and F value. Applying this model to the named entity recognition of Chinese medicine texts, the best results can be obtained in a small sample of data.