中国医药科学2025，Vol.15Issue(18)：99-102,146,5.DOI:10.20116/j.issn2095-0616.2025.18.19
基于机器学习算法的脑卒中疾病风险预测模型
Stroke disease risk prediction model based on machine learning algorithms
摘要
Abstract
Objective To construct a stroke risk prediction model based on machine learning techniques and identify key influencing factors of stroke onset.Methods Data from inpatients at Hongqi Hospital Affiliated to Mudanjiang Medical University(5000 samples)between March 2022 and December 2024 were collected,including 12 features such as gender,age,and hypertension history.After data cleaning,feature engineering,missing value imputation,and standardization,the Synthetic Minority Over-sampling Technique(SMOTE)was applied to balance the dataset.Three machine learning algorithms-logistic regression(LR),random forest(RF),and support vector machine(SVM)-were employed to develop prediction models.Model performance was evaluated using metrics including AUC,accuracy,and sensitivity.Results The Random Forest model achieved optimal performance(AUC=0.78,accuracy=71%),with significantly higher specificity(70%)and positive predictive value(71%)compared to other models.Interpretability analysis based on SHAP indicates that age and average blood glucose level are the most critical influencing factors for the onset of stroke.Conclusion This study innovatively integrates the Shapash framework for model visualization and interpretation,providing a reliable tool for early clinical identification of high-risk populations and personalized intervention strategies.The results demonstrate that the model combines high predictive performance with clinical interpretability,offering significant insights for optimizing stroke prevention and treatment systems.关键词
机器学习/脑卒中/风险预测/随机森林/可解释性分析Key words
Machine learning/Stroke/Risk prediction/Random forest/Explainability analysis分类
医药卫生引用本文复制引用
冯莹,张天宁,李明旭,杜禹辛,于成龙..基于机器学习算法的脑卒中疾病风险预测模型[J].中国医药科学,2025,15(18):99-102,146,5.基金项目
黑龙江省省属高等学校基本科研业务费科研项目(2023-KYYWF-0962). （2023-KYYWF-0962）
