Screening of Syndrome Elements and Construction of Diagnostic Model of Hyperactivity of Liver-Yang Syndrome Based on Machine Learning

Objective: Based on a variety of machine learning algorithms, screen the hyperactivity syndrome e elements of liver yang syndrome. Methods: Using the classification of hyperactivity of liver yang syndrome standards and searching for relevant standards of hyperactivity of liver yang syndrome, The study constructed a matrix data set containing the syndrome elements of hyperactivity of liver yang syndrome and screened out key syndrome elements with the help of three machine learning algorithms, including XGBoost, SVM, and RF. Based on the results of the expert questionnaire, the key syndrome elements of liver yang syndrome are derived. Results: A total of 10 key elements were screened out through 3 machine learning algorithms. Combined with the results of multi region, multi field and multi-level research, the core syndrome elements of irritability, soreness and softness of the waist and knees, stringy pulse and red tongue were obtained. Conclusion: There is still a need to expand clinical samples and continuously optimize the diagnosis model of hyperactivity of liver-Yang syndrome based on expert opinions because of this study, as it provides a data base for the digitalization and objectification of TCM syndrome differentiation.