南疆地区女性压力性尿失禁的预测模型构建  

Construction of a prediction model of female stress urinary incontinence in Southern Xinjiang

作　　者：王元毅 曾凯[1] 郝雨[1] 李晨[1] 倪钊[1] 李强[1] 王勤章[1] WANG Yuanyi;ZENG Kai;HAO Yu;LI Chen;NI Zhao;LI Qiang;WANG Qinzhang(Department of Urology,The First Affiliated Hospital of Medical College,Shihezi University,Shihezi 832000,China)

机构地区：[1]石河子大学医学院第一附属医院泌尿外科,新疆石河子832000

出　　处：《现代泌尿外科杂志》2021年第9期735-739,共5页Journal of Modern Urology

基　　金：兵团科技攻关项目(No.2018AB023);兵团科技计划项目(No.2021AB032)。

摘　　要：目的依据南疆地区女性压力性尿失禁(SUI)相关数据,建立多层感知器(MLP)神经网络预测模型,为女性SUI的早期干预提供参考。方法通过对南疆地区女性盆底功能障碍性疾病进行流行病学调查,获取其中SUI相关数据,同时对我国女性SUI的危险因素进行Meta分析,获得影响SUI的相关因素,并建立MLP神经网络预测模型,同时对该模型进行测试及验证。结果MLP神经网络预测模型的准确度为86.8%,通过验证提示灵敏度为93.1%、特异度为80.43%、准确度为85.33%,曲线下面积(AUC)为0.924,基尼(Gini)系数为0.848。结论该模型具有良好的预测效果,能够对预防及干预SUI的发生提供参考。Objective To establish a multi-layer perceptron(MLP)neural network prediction model based on the data of female stress urinary incontinence(SUI)in Southern Xinjiang,in order to provide reference for the early intervention of SUI.Methods An epidemiological investigation was conducted to obtain data of SUI.The influencing factors of SUI were analyzed with a Meta-analysis.An MLP neural network prediction model was established and verified.Results The accuracy of the prediction model was 86.8%.Verification indicated that the accuracy,sensitivity and specificity of the prediction model were 85.33%,93.1%and 80.43%,respectively.The AUC was 0.924,and the Gini index was 0.848.Conclusion The model has good predictive effects and can provide reference for prevention and intervention of SUI.

关 键 词：新疆南疆 女性 压力性尿失禁 神经网络 预测模型 

分 类 号：R699.7[医药卫生—泌尿科学]