摘要
Abstract
Corneal shaping lens is the most widely used optical means to control the development of myopia in adolescents in clinical practice,however,there are problems such as inefficiency in the fitting process of traditional corneal shaping lens.This arti-cle explores the intelligent fitting of four important parameters,namely AC,reduction,diameter and CP,by applying machine learning algorithm to assist Mondial keratomileusis.Linear regression,decision tree,ETR,Adaboost and GDBT algorithms are used to construct prediction models for keratoconus and the importance scores of each factor indicator are given in conjunction with the model prediction results.The ETR,Adaboost and GBDT regression models achieve good results in predicting the parameters AC,drop and diameter,respectively,with MSEs of 0.146,0.088 and 0.026 for the test set.The accuracy of the test set is 0.961.The machine learning models based on ETR,Adaboost and GBDT can predict the important parameters of keratoconus better,which can largely improve the efficiency of keratoconus fitting.关键词
角膜塑形镜/ETR/Adaboost/GBDTKey words
orthokeratology lens/ETR/Adaboost/GBDT分类
信息技术与安全科学