医学信息2025，Vol.38Issue(22)：10-17,8.DOI:10.3969/j.issn.1006-1959.2025.22.002
用机器学习算法建立膜性肾病的鉴别诊断模型
Establishing a Differential Diagnosis Model for Membranous Nephropathy by Machine Learning Algorithms
摘要
Abstract
Objective To explore the most suitable classification algorithm for the identification of primary membranous nephropathy(PMN),and to provide data reference for PMN diagnosis research.Methods A total of 500 patients diagnosed by renal pathology in Luohe Central Hospital from June 2019 to June 2021 were selected.All patients were confirmed as primary glomerular disease by renal biopsy,including 322 cases of PMN and 178 cases of non-PMN.Decision tree,random forest,support vector machine and limit gradient boosting algorithm(Xgboost)were used to establish the differential diagnosis model of PMN and non-PMN.The best performance of the model was evaluated according to the true positive rate,true negative rate,false positive rate,false negative rate,accuracy,and area under the ROC curve(AUC)of the subjects.Results The F1 score of Xgboost model in PMN was the highest(0.95).Correlation analysis showed that anti-PLA2R,UPQ/24h,age,C3,C4 and gender were positively correlated with PMN.The Xgboost model based on the above evaluation indexes had the highest efficiency in diagnosing PMN,and its sensitivity and specificity were 92.00%and 96.00%,respectively.Conclusion The differential diagnosis model of PMN was successfully established.The Xgboost model has the best effectiveness and can be used for clinical diagnosis of PMN.关键词
原发型膜性肾病/机器学习算法/决策树/随机森林/极限梯度提升算法Key words
Primary membranous nephropathy/Machine learning algorithms/Decision tree/Random forest/The limit gradient lifting algorithm分类
医药卫生引用本文复制引用
曹尚美,陈泊霖,付秀虹..用机器学习算法建立膜性肾病的鉴别诊断模型[J].医学信息,2025,38(22):10-17,8.基金项目
1.河南省自然科学基金(编号:2223000420247) （编号:2223000420247）
2.2023年河南省博士后科研资助项目(编号:HN2024098) （编号:HN2024098）
3.中央引导地方项目(编号:Z20221343023) （编号:Z20221343023）
4.河南省医学科技攻关项目(编号:LHGJ20221031) （编号:LHGJ20221031）
