Sasang constitutional medicine is a unique concept in Korean medicine that can provide valuable insights into personalized healthcare and disease treatment. In this study, we combined network analysis and information entropy to systematically investigate the related information of Sasang constitutional (SC) types. A feature network was constructed using SC type and clinical information. The SC type-associated features and feature classes were identified using statistical analysis and entropy ranking. The patient network was constructed based on SC-type-associated features. We found that the feature network was closely connected within the features of the same classes and between several feature class pairs, including the symptom class. Most of the separation values between the feature classes, including the symptom class, were negative. In addition, we found 42 clinical features related to the SC type, and two important classes -personality and cold/heat- that increase the entropy ranking of the SC type. In the patient network, we found sparsely connected modules between SC types and a positive separation value between the Taeeumin-Soeumin and Taeeumin-Soyangin pairs. Our data-driven approach provides a deeper understanding of modernized forms of SC types and suggests that SC type is a practically useful concept for stratified healthcare and personalized medicine.
Keywords: Sasang constitutional medicine; entropy; network-based analytics; personalized medicine.