摘要
为提升冠状动脉粥样硬化性心脏病的防治水平,赋能智慧医疗,以心内科医疗信息化数据为对象,对冠状动脉粥样硬化性心脏病早期预测方法展开研究。阐述了从医疗信息系统中采集模型所需样本的数据筛选、数据清洗、文本预处理、文本表示以及特征归一化方法流程,提出了一种基于粒子群优化的随机森林预测方法。该方法以k折交叉验证平均准确率为目标函数,能够自适应优化随机森林的模型参数,从而提升模型的分类能力。实验结果表明,在真实世界冠心病数据的预测上,该方法具有较高的预测精度,准确率为87%,灵敏度为87%,特异度为88%,AUC为0.91,较未优化的随机森林具有明显的提升。因此,对冠状动脉粥样硬化性心脏病的早期预测应用具有一定的参考价值。
To raise the prevention and treatment level of coronary atherosclerotic heart disease(CAHD)and empower intelligent medical treatment,the early prediction method of CAHD was studied with the medical information data of Department of Cardiology as the objects.This paper sets forth the methods and processes of data screening,data cleaning,text preprocessing,text representation and feature normalization for collecting samples required by the model from the medical information system,and proposes a random forest prediction method based on particle swarm optimization.With the average accuracy rate of k-fold cross validation as the objective function,this method can optimize the model parameters of random forest adaptively,thus improving the classification ability of the model.The experimental results show that this method has a high prediction accuracy of 87%,the sensitivity of 87%,the specificity of 88%,and AUC of 0.91 in the prediction of real CAHD data,which are significantly higher than those of the nonoptimized random forest.Therefore,it has a certain reference value for the early prediction of CAHD.
作者
韩刚
卢鹏飞
陈珊黎
邵维君
贾红岩
郑涛
Han Gang;Lu Pengfei;Chen Shanli;Shao Weijun;Jia Hongyan;Zheng Tao(Information Center,Renji Hospital,Shanghai Jiaotong University School of Medicine,Shanghai 200127,China;Wonders Information Co.,Ltd.)
出处
《中国数字医学》
2022年第4期56-61,共6页
China Digital Medicine
基金
上海市信息化发展专项资金项目-面向仁济医院医联体的专病临床科研智能辅助决策平台建设(201901007)。
关键词
粒子群算法
随机森林
冠状动脉粥样硬化性心脏病
智慧医疗
Particle swarm optimization
Random forest
Coronary atherosclerotic heart disease
Intelligent medical treatment
