中文电子病历命名实体识别算法BLF-MarkBERT

Chinese electronic health record named entity recognition algorithm BLF⁃MarkBERT

随着深度学习技术的发展,中文命名实体识别在各个领域取得了显著进展,特别是在中文电子病历领域,它成为了医学信息管理领域的重要任务。中文电子病历命名实体识别从电子病历中自动识别和分类命名实体,提高了医学信息管理效率和临床决策支持,促进了医学智能信息化发展。为进一步提升效果,对MarkBERT方法进行研究,在其基础上改进并实现了一种融合双向长短时记忆网络和解码方式的深度学习模型BLF-MarkBERT。在CCKS2019数据集上的实验结果表明,BLF-MarkBERT在准确率P、召回率R和F1分数这三个评估指标上均优于对比算法,表明了该模型的优越性。

With the development of deep learning technology,Chinese named entity recognition has made significant progress in various fields,especially in the field of Chinese electronic medical records,where it has become an important task in the field of medical information management.Chinese electronic medical record named entity recognition automatically identifies and classifies named entities from electronic medical records,improves medical information management efficiency and clinical decision support,and promotes the development of medical intelligent informatization.In order to further improve its effect,this paper improves and implements a deep learning model BLF‑MarkBERT based on the MarkBERT method that integrates the bidirectional long short-term memory network and decoding method.Experimental results on the CCKS2019 data set show that BLF‑MarkBERT is better than the comparison algorithm in the three evaluation indicators of Precision(P),Recall(R)and F1‑score,indicating the superiority of the model.