Related Experiment Video
Updated: Jun 17, 2026
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
Diagnostic accuracy of pattern differentiation algorithm based on Chinese medicine theory: a stochastic simulation
1Department of Rehabilitation Science, Centro Universitário Augusto Motta, Rio de Janeiro, BR, Brazil. arthur_sf@ig.com.br
Chinese Medicine
|December 23, 2009
Summary
This study shows that using both explained and optimal available information improves diagnostic accuracy for Zang-fu pattern differentiation in Chinese medicine. Incorporating a cutoff value enhances the performance of the pattern differentiation algorithm (PDA).
Area of Science:
- Traditional Chinese Medicine
- Diagnostic Algorithms
- Computational Medicine
Background:
- Clinical practice of Chinese medicine often lacks sufficient data for accurate Zang-fu pattern differentiation.
- This study investigates the impact of varying information quantities on diagnostic accuracy.
Purpose of the Study:
- To evaluate the effect of information amount on the diagnostic accuracy of a pattern differentiation algorithm (PDA).
- To determine optimal criteria for Zang-fu pattern differentiation using computational methods.
Main Methods:
- A dataset of 69 Zang-fu patterns with manifestations from the Four Examinations (inspection, auscultation/olfaction, inquiry, palpation) was used.
- Stochastic simulation varied the amount of available information (N(%)) to test the PDA's diagnostic performance.
- Accuracy, sensitivity, specificity, and predictive values were calculated for different combinations of diagnostic information.
Main Results:
- The Four Examinations provided the highest accuracy (94.7%) with the smallest information cutoff (N(%)=28.5%).
- Accuracy curves showed an optimal value dependent on the N(%-cutoff).
- Using both explained information (F(%)) and optimal available information cutoff (N(%-cutoff)) significantly improved diagnostic metrics.
Conclusions:
- Pattern differentiation is more accurate when using both explained (F(%)) and optimal available information (N(%-cutoff)) compared to using them without a cutoff.
- The study recommends using both F(%) and N(%-cutoff) as objective criteria for the PDA in Zang-fu pattern differentiation.
