临床医学工程2025，Vol.32Issue(2)：113-116,4.DOI:10.3969/j.issn.1674-4659.2025.02.0113
机器学习模型在角膜塑形镜验配中的价值研究
Study on the Value of Machine Learning Models in Orthokeratology Fitting
摘要
Abstract
Objective To explore the practical value of machine learning models in orthokeratology fitting,and analyze the prediction accuracy.Methods The clinical data of 644 patients who were fitted with orthokeratology in our hospital from October 2021 to June 2023 were retrospectively analyzed,and five machine learning models(XGBoost,RF,SVR,ridge regression,and stacking models)were established on this basis.The prediction results of each model were compared with the parameters of the ordered lens.Results The prediction results of each model were highly consistent with the parameters of the ordered lens.Compared with other models,the stacking model had a higher accuracy in the prediction of BC,RC,AC1,and AC2 with R2 of 0.935,0.990,0.945,and 0.886,respectively.Combined with evaluation indicators such as MSE,stacking model had a smaller error.Conclusions Among the machine learning models,stacking model has higher prediction accuracy on orthokeratology fitting,which can be used as an auxiliary fitting method for orthokeratology.关键词
角膜塑形镜/机器学习/模型准确性/近视控制Key words
Orthokeratology/Machine learning/Model accuracy/Myopia control分类
医药卫生引用本文复制引用
黎崎鸿,王武..机器学习模型在角膜塑形镜验配中的价值研究[J].临床医学工程,2025,32(2):113-116,4.基金项目
湖南省自然科学基金项目(2023JJ70037) （2023JJ70037）
