Named Entity Recognition in Electronic Medical Records Based on Transfer Learning
ID：54
Submission ID：59 View Protection：ATTENDEE
Updated Time：2025-10-11 22:35:10 Hits：75
Poster Presentation
Start Time：2025-11-09 09:04 （Asia/Shanghai）
Duration：1min
Session：[P] Poster presentation » [P6] 6.AI-driven technology
No files
Abstract
To address the challenges of scarce labeled data and cross-disease knowledge transfer in named entity recognition (NER) tasks for Chinese electronic medical records, this paper proposes a transfer learning method based on the BERT-BiLSTM-CRF model to explore cross-disease knowledge transfer strategies. By comparing experimental results in non-transfer and transfer learning scenarios, we systematically explore the impact of the number of target domain samples and the ratio of source and target domain data on model performance. Baseline model experiments show that differences in data distribution have a significant impact on entity recognition performance; after introducing transfer learning, the model's recognition performance in the target domain (especially in small sample scenarios) is significantly improved. This research provides an effective technical solution for low-resource medical text processing.
Keywords
BERT-BiLSTM-CRF,Chinese electronic medical records,named entity recognition,transfer learning
Speaker
Comment submit