Normalizing Symptom Names in Traditional Chinese Medicine Records
Author Information
Author(s): Wang Yaqiang, Yu Zhonghua, Jiang Yongguang, Xu Kaikuo, Chen Xia
Primary Institution: Sichuan University
Hypothesis
Can automatic normalization of symptom names improve data mining in traditional Chinese medicine?
Conclusion
The study confirms that hybrid similarity metrics are more effective for normalizing symptom names in clinical records of TCM than using either literal or remedy-based metrics alone.
Supporting Evidence
- The hybrid similarity metric achieved the highest F-Measure of 65.62%.
- The study found that 78.41% of standard symptom names have multiple synonyms.
- The hybrid metrics outperformed both literal and remedy-based metrics in accuracy.
Takeaway
This study helps computers understand different names for the same symptoms in traditional Chinese medicine, making it easier to analyze medical records.
Methodology
The study developed algorithms using hybrid similarity metrics that combine literal and remedy-based similarities to normalize symptom names.
Limitations
The study primarily focuses on symptom names and may not address other aspects of clinical records.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website