基于CNN与多种机器学习算法投票的心脏病预测模型
计算机时代 毕吉霖 肖作文 何红宇
中图分类号：TP39 文献标识码：A文章编号：1006-8228（2025）10-25-06
Heart Disease Prediction Model Based on CNN and Voting of Multiple Machine Learning Algorithms
Bi Jilin，Xiao Zuowen，He Hongyu
（SchoolofElectronicInformationandElectricalEnginering，YangtzeUniversityJingzhou，Hubei4332ia）
Abstract：Predictionofheartdiseaseusingmachinelearningmodelscanasistdoctorsindiagnosis；therefore，itisimportantto carryoutresearchonsuchmodels.Inthispaper，apredictionmodelbasedonconvolutionalneuralnetwork（CNN）andvotingof multiplemachinelearingalgoritmsisproposed.Firstthetabulardataarepreprocessed;then，featurecombinationandextraction areperformedbytheCNNmoduletoobtainfeaturevectors;next，thefeaturevectorsareinputintofourmachinelearning algorithms，namelyKNLogisticRegresionAdaboostandSVMandsoftvotingisapliedtoobtainthepreditiosults. Comparedwithmachinelearningpredictionmodelswithoutapplyingvoting，ourmodelshowsbeterresultsinheartdisease prediction：accuracyrateof0.93，precisionrateof0.945，recallrateof0.862，andAUCvalueof0.970.Thevoting-basedmodel ofmultiplemachinelearningalgorithmsproposedinthispapercanprovidedoctorswithmoreeffectivesupportinheartdisease diagnosis.
KeyWords：HeartDiseasePrediction;MachineLearning;SoftVoting;ConvolutionalNeuralNetwork;FeatureCombination
0引言
目前，我国心脏病患者数量持续增加、患者年龄逐渐下降，因此，心脏病已受到越来越多的关注[。（剩余7672字）