护理与康复2025，Vol.24Issue(6)：1-6,6.DOI:10.3969/j.issn.1671-9875.2025.06.001
基于3种机器学习算法构建脑卒中患者隐性误吸风险预测模型
Construction of a risk prediction model for silent aspiration in stroke patients based on three machine learning algorithms
摘要
Abstract
Objective To develop a risk prediction model for silent aspiration in stroke patients using three machine learning algorithms.Methods A retrospective study was conducted on 550 stroke patients with dysphagia admitted to the First Affiliated Hospital of Wenzhou Medical University from January 2020 to March 2024.Influencing factors for silent as-piration were identified based on the biopsychosocial model.Python 3.11 was used to analyze three machine learning algo-rithms—random forest,support vector machine,and XGBoost—to construct risk prediction model.The predictive perform-ance of the models was evaluated through the receiver operating characteristic(ROC)curve,the area under the ROC curve(AUC),accuracy,precision,recall rate,and F1 score.The optimal model was selected by comparing these metrics.Results The incidence of silent aspiration post-stroke was 57.09%.Five variables were ultimately included in the prediction model,swallo-wing function,patients'perception of silent aspiration risk,food texture,number of comorbidities,and number of strokes.The XGBoost model demonstrated the highest accuracy(60.91%)in the validation set,with an AUC of 0.71,recall rate of 74.11%,and precision of 62.07%.Conclusion The risk prediction model for silent aspiration in stroke patients based on three machine learning algorithms can effectively identify key influencing factors.Applying this model is helpful for improving the early recognition rate of silent aspiration.关键词
脑卒中/隐性误吸/机器学习/风险预测模型Key words
stroke/silent aspiration/machine learning/risk prediction model分类
医药卫生引用本文复制引用
洪显钗,舒美春,林碎丽,翁一心,鲍少蕊..基于3种机器学习算法构建脑卒中患者隐性误吸风险预测模型[J].护理与康复,2025,24(6):1-6,6.基金项目
温州市基础性科研项目,编号Y2023054 （）
