基于BERT-BILSTM-CRF的慢性支气管炎中医医案实体识别

Entity Recognition of Traditional Chinese Medical Cases of Chronic Bronchitis Based on BERT-BILSTM-CRF

随着现代信息技术的飞速发展,人类社会开始进入大数据时代,如何高效快捷地从海量的中医医案文本数据中挖掘出我们所需要的信息,从而更好地应用于临床工作,是目前亟待解决的问题。通过实验对慢性支气管炎中医医案进行研究,分析BERT、BILSTM、BILSTM-CRF和BERT-BILSTM-CRF四种模型的实体识别效果,结果表明,相比于其他模型,采用BERT-BILSTM-CRF模型可以更加准确有效地识别出慢性支气管炎中医医案的实体类别,其F1、Precision和Recall均优于其他模型。

With the rapid development of modern information technology,human society has begun to enter the era of big data.How to efficiently and quickly mine the information we need from the massive text data of traditional Chinese medicalcases,so as to better apply them to clinical work,which is an urgent problem to be solved at present.Based on the experimental study of traditional Chinese medicalcases of chronic bronchitis,the entity recognition effects of four models,BERT,BILSTM,BILSTM-CRF and BERT-BILSTMCRF,are analyzed.The results show that compared with other models,the BERT-BILSTM-CRF model can more accurately and effectively identify the entity categories of traditional Chinese medicalcases of chronic bronchitis,and its F1,Precision and Recall are all better than that of other models.