Chinese clinical named entity recognition with variant neural structures based on BERT methods.

Area of Science:

• Natural Language Processing
• Medical Informatics
• Artificial Intelligence

Background:

• Clinical Named Entity Recognition (CNER) is crucial for processing electronic medical records.
• Deep learning models show promise but require extensive labeled data, which is scarce for Chinese clinical text.
• Existing methods face challenges in accurately identifying and classifying clinical terms in Chinese medical records.

Purpose of the Study:

• To develop a high-performance Chinese CNER model addressing data scarcity.
• To leverage unlabeled Chinese clinical records for domain-specific knowledge.
• To improve the accuracy of clinical term identification and classification.

Main Methods:

• Pre-training a BERT model on unlabeled Chinese clinical records.
• Utilizing Long Short-Term Memory (LSTM) and Conditional Random Field (CRF) layers for feature extraction and tag decoding.
• Incorporating a novel strategy for dictionary features and radical features of Chinese characters.

Main Results:

• The ensemble model achieved a strict F1 score of 89.56% on the CCKS-2018 dataset.
• The model attained an F1 score of 91.60% on the CCKS-2017 dataset.
• Outperformed existing state-of-the-art models on both datasets.

Conclusions:

• The proposed ensemble model effectively addresses the challenges of Chinese CNER, particularly data scarcity.
• Leveraging domain-specific knowledge through pre-training and incorporating linguistic features significantly enhances model performance.
• This approach offers a robust solution for clinical named entity recognition in Chinese medical records.