Related Experiment Video
Updated: Oct 16, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.7K
Natural Language Processing Algorithms for Normalizing Expressions of Synonymous Symptoms in Traditional Chinese
Lu Zhou1, Shuangqiao Liu1, Caiyan Li1
1Beijing University of Chinese Medicine, School of Traditional Chinese Medicine, Beijing 100029, China.
Evidence-Based Complementary and Alternative Medicine : Ecam
|October 21, 2021
Summary
This study developed a BERT-Classification model to normalize Traditional Chinese Medicine (TCM) symptoms. The model effectively unifies synonymous TCM symptoms, improving data mining for TCM modernization.
Area of Science:
- Computational linguistics
- Medical informatics
- Traditional Chinese Medicine (TCM)
Background:
- Modernizing Traditional Chinese Medicine (TCM) requires systematic data mining from medical records.
- TCM faces challenges due to synonymous symptoms with different expressions, hindering data analysis.
- Natural language processing (NLP) offers a solution for normalizing TCM symptoms.
Purpose of the Study:
- To construct and compare NLP models for normalizing TCM synonymous symptoms.
- To identify the highest-performing model for unifying TCM symptom expressions.
Main Methods:
- Developed four NLP models: Bi-LSTM (sequence generation and classification) and BERT (BERT-UniLM and BERT-Classification).
- Evaluated models using accuracy, recall, precision, and F1-score.
- Compared performance of Bi-LSTM and BERT-based models.
Main Results:
- The BERT-Classification model demonstrated superior performance across all four evaluation metrics.
- BERT-Classification outperformed Bi-LSTM and BERT-UniLM models in TCM symptom normalization.
Conclusions:
- The BERT-Classification model is highly effective for normalizing expressions of TCM synonymous symptoms.
- This advancement supports systematic data mining and modernization of TCM.
