摘要
目的评价反向传播(back propagation,BP)神经网络模型在子宫肌瘤危险因素分析中的应用价值.方法采用分层随机抽样的方法对沈阳地区.1260名妇女的子宫肌瘤患病状况进行问卷调查,对113例子宫肌瘤病例采用1:2配比病例对照研究,利用MATLAB 6.5软件的神经网络工具箱构建BP神经网络模型,训练与模拟网络,分析子宫肌瘤各种可能危险因素的平均影响值(meanimportant value,MIV),并与多因素条件logistic回归模型的分析结果相比较,用对数线性模型分析因子间的交互作用.结果BP神经网络分析结果显示,影响子宫肌瘤发病主要危险因素为月经初潮年龄延迟、母亲或姐妹患子宫肌瘤、宫颈炎、月经紊乱、人工流产史、盆腔炎、口服避孕药、阴道炎,其对应的MIV值分别为-0.0405、0.0361、0.0162、0.0143、0.0135、0.0117、0.0094、0.0087;比较BP神经网络输出的MIV与多因素条件logistic回归分析结果,发现两者主要发病危险因素排列顺序基本一致,但存在一些差异,经对数线性模型分析发现人工流产史可能是子宫肌瘤发病的一个重要的协同变量.结论与传统数学模型相比,BP神经网络能较好地处理数据协变量间的交互作用,是一种很好的危险因素分析方法.
Objective To evaluate the value of a back propagation(BP)network in determining the risk factors of uterine myomas.M lethods Using stratified randomized sampling method,1260 women were surveyed by questionnaire.1:2 matched case contol study was conducted in 113 cases of uterine myomas.Neural network tols box of Software MATLAB 6.5 was used to train and simulate BP arificial network.The mean impact value(MIV)for each input variables was analyzed,and Wa8 compared with multiple logistic regression analysis and log-linear model for interaction between factors.Results BP artificial neural analysis showed that the leading risk factors for uterine myomas were delayed menstruation,family history of uterine myomas,cervicitis,menstrual disorder,induced abortion,pelvic inlammatory,oral contraceptive medication,and elytritis,with mean impact value-0.0405,0.0361,0.0162,0.0143,0.0135,0.0117,0.0094,0.0087,respectively.Both BP artificial neural and logistic egression analysis showed that the sequence of leading risk factors were similar in the whole,but there were some diferences observed,induced abortion was proved to be an important cooperation variable through logline model analysis respectively.Conclusion Compared to the conventional statistics method,BP artificial neural network could deal with the interaction between covariables preferably,thus provided a powerful method to risk factor analysis.
作者
王玮
许伟
周宝森
WANG Wei;XU Wei;ZHOU Bao-sen(Department of Epidemiology,China Medical University,Shenyang 1000,China;不详)
出处
《中华预防医学杂志》
CAS
CSCD
北大核心
2007年第S01期94-97,共4页
Chinese Journal of Preventive Medicine
基金
国家自然科学基金资助项目(39770675)
