Modification of the Integrated Sasang Constitutional Diagnostic Model

Area of Science:

• Traditional Korean Medicine
• Biomedical Informatics
• Computational Statistics

Background:

• The Sasang constitutional diagnostic method, a traditional Korean medicine approach, faces quantification challenges.
• A previous objective physical diagnostic model (2012) requires revision due to identified issues.

Purpose of the Study:

• To propose methodological improvements for the Sasang constitutional diagnostic model.
• To enhance the accuracy and comprehensiveness of Sasang constitutional type classification.

Main Methods:

• Variable selection and Least Absolute Shrinkage and Selection Operator (LASSO) for multicollinearity reduction.
• Development of individual diagnostic models considering Sasang constitutional type proportions and age.
• Construction of an integrated model using explanatory variables from individual models.

Main Results:

• Significant improvements in sensitivity for the male SY type (36.4% to 62.0%) and female SE type (43.7% to 64.5%).
• Enhanced accuracy in classifying Sasang constitutional types compared to the previous model.

Conclusions:

• The proposed integrated diagnostic model offers a more accurate and objective approach to Sasang constitutional diagnosis.
• These algorithmic advancements are expected to benefit the scientific development of Sasang constitutional medicine and other traditional diagnostic systems.