中国医疗设备2025，Vol.40Issue(8)：38-42,76,6.DOI:10.3969/j.issn.1674-1633.20241191
基于PCA和随机森林的医院分诊模型研究
Research on Hospital Triage Model Based on PCA and Random Forest
摘要
Abstract
Objective To address the problems of low triage efficiency,inaccurate triage and privacy leakage in traditional triage methods,to establish an intelligent triage model,so as to enhance the efficiency,accuracy and privacy protection ability of triage.Methods The triage algorithm was designed by integrating principal component analysis and random forest algorithm,and differential privacy and homomorphic encryption technology were introduced to construct a dual-encrypted hospital emergency triage model.Results The triage algorithm designed in the study achieved convergence after the 4th and 17th iterations respectively in the tests on two datasets.The fitting degrees were 96.8%and 98.4%respectively.It outperformed other algorithms in terms of computational efficiency and triage accuracy.The average calculation time of the research model was 0.43 s.The triage accuracy was higher than that of other triage models within the range of the total budget value for privacy protection ranging from 0.01 to 5.00.There were statistically significant differences in the waiting time for consultation,triage accuracy and readmission rate before and after the application of the emergency triage model(P<0.05),indicating that the model in this study significantly improved the triage capacity of the hospital.Conclusion Under the condition of safeguarding patients'privacy,the model in this study can efficiently and accurately complete triage and improve the utilization rate of medical resources.关键词
主成分分析法/随机森林(RF)/分诊模型/差分隐私(DP)/同态加密(HE)/隐私泄露Key words
principal component analysis(PCA)method/random forest(RF)/triage model/differential privacy(DP)/homomorphic encryption(HE)/privacy leakage分类
医药卫生引用本文复制引用
李璇..基于PCA和随机森林的医院分诊模型研究[J].中国医疗设备,2025,40(8):38-42,76,6.基金项目
江苏省卫生经济学会2024年度卫生经济研究课题(JSWSJJ202409). （JSWSJJ202409）
