Objective To construct a Chinese medicine(CM)pattern differentiation model for systemic lupus erythematosus(SLE)using the XGBoost algorithm and explore the feasibility of applying the XGBoost model for CM pattern classification.Methods Eligible cases were collected through a questionnaire survey to establish a SLE dataset.An XGBoost-based SLE CM pattern differentiation model was developed,and the random forest(RF)algorithm was used as a control for accuracy comparison.Results A total of 400 SLE patients were included in this study,including 33 males and 367 females.The top three CM patterns for SLE patients were yang deficiency of the spleen and kidney pattern,yin deficiency-induced internal heat pattern,and wind dampness and heat impediment pattern.The classification indicators and performance curve scores of the XGBoost algorithm model were overall superior to those of the RF algorithm.Conclusion XGBoost algorithm demonstrates high accuracy in CM pattern modeling and can be used for classification research in CM pattern studies.
关键词
系统性红斑狼疮/XGBoost算法/随机森林算法/中医证候
Key words
systemic lupus erythematosus/XGBoost algorithm/random forest algorithm/Chinese medicine pattern
维普
万方数据