A New Method for Medical Triage Using Text Classification
Author Information
Author(s): Xiao Xiao, Wang Shuqin, Jiang Feng, Qi Tingyue, Wang Wei
Primary Institution: The Affiliated Hospital of Yangzhou University
Hypothesis
Can a cross-domain text classification method based on prompt-tuning effectively alleviate the medical strain in triage systems?
Conclusion
The proposed method outperforms existing state-of-the-art methods for medical triage recommendation.
Supporting Evidence
- The proposed method achieved state-of-the-art performance on Chinese Triage datasets.
- Extensive experiments demonstrated the effectiveness of the prompt-tuning approach.
- The method can classify patient inquiries with only a small number of labeled instances.
Takeaway
This study created a smart way to help patients find the right doctor by using computer programs that understand their questions about symptoms.
Methodology
The study used a cross-domain text classification method based on prompt-tuning to classify patient inquiries into medical departments.
Potential Biases
Potential biases may arise from the reliance on specific templates and the quality of the training data.
Limitations
The method may still require some labeled data and may not fully eliminate semantic information loss.
Participant Demographics
The study focused on Chinese patients seeking medical advice.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website