基于Super Learner集成学习模型对脑卒中的预测研究  

A Study of Stroke PredictionBased on Super Learner Integrated Learning Model

作　　者：孙浩浩 章剑林[1] 张子蓥 黄可欣 SUN Haohao;ZHANG Jianlin;ZHANG Ziying;HUANG Kexin(Alibaba Business School,Hangzhou Normal University,Hangzhou 311121,China)

机构地区：[1]杭州师范大学阿里巴巴商学院,浙江杭州311121

出　　处：《杭州师范大学学报（自然科学版）》2023年第6期590-597,610,共9页Journal of Hangzhou Normal University(Natural Science Edition)

摘　　要：使用索马里医院提供的脑卒中患者数据集,通过四分位距(interquartile range, IQR)方法和合成少数类过采样技术(synthetic minority oversampling technique, SMOTE)算法进行数据预处理,采用特征工程中的嵌入式方法对数据集进行特征分析,确定脑卒中诱发因素.以随机森林(random forest, RF)、极端梯度提升(extreme gradient boosting, XGB)和自适应提升(adaptive boosting, AdB)算法为第一层,高斯朴素贝叶斯(Gaussian naive bayes, GaNB)和支持向量机(support vector machine, SVM)为第二层,逻辑回归(logistic regression, LR)为元学习器构建超级学习者(super learner, SL)集成学习模型.仿真实验结果表明,相较于6种基础算法,SL模型预测效果最优,可为脑卒中的预测分析提供新的选择.Using the dataset of stroke patients provided by Somali hospitals,the data were preprocessed by interquartile range(IQR)method and synthetic minority oversampling technique(SMOTE)algorithm.The embedded method in feature engineering was used to characterize the dataset and identify the stroke triggers.Random forest(RF),extreme gradient boosting(XGB)and adaptive boosting(AdB)algorithms were used as the first layer,Gaussian naive bayes(GaNB)and support vector machine(SVM)algorithms were used as the second layer,and logistic regression(LR)was used to construct the super learner(SL)integrated learning model for meta-learning.The results of simulation experiments showed that the SL model had the best prediction performance compared with the other six basic algorithms,which could provide a new choice for stroke prediction and analysis.

关 键 词：脑卒中 特征分析 super learner 机器学习 预测 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]