基于支持向量机的ICU中急性低血压预测模型研究  

Study on Predicting Model for Acute Hypotensive Episodes in ICU Based on Support Vector Machine

作　　者：赖丽娟[1] 王志刚[1] 吴效明[1] 熊冬生[1] 

机构地区：[1]华南理工大学生物医学工程系,广州510006

出　　处：《生物医学工程学杂志》2011年第3期451-455,共5页Journal of Biomedical Engineering

基　　金：广东省科技计划项目资助(2009B030801004)

摘　　要：在重症监护室(ICU)的监护中,急性低血压(AHE)的发生严重威胁着患者的生命安全,临床上主要依靠医生的经验处置。本文运用医学信息学的理论,研究一种ICU中AHE发生的预测模型。利用ICU监护中血压变化的连续记录数据,分析发生与未发生AHE两者间平均动脉压(MAP)信号的变化趋势与特点,基于统计学习理论的支持向量机(SVM)方法,选取中位数、平均值等统计特征参数用于学习和训练,建立分类预测模型。在此基础上,对不同核函数构成的分类器和预测算法进行了比较分析。实验验证,本方法能够达到比较好的分类预测效果,有利于AHE发生的提前预测。The occurrence of acute hypotensive episodes（AHE） in intensive care units（ICU） seriously endangers the lives of patients,and the treatment is mainly depended on the expert experience of doctors.In this paper,a model for predicting the occurrence of AHE in ICU has been developed using the theory of medical Informatics.We analyzed the trend and characteristics of the mean arterial blood pressure（MAP） between the patients who were suffering AHE and those who were not,and extracted the median,mean and other statistical parameters for learning and training based on support vector machine（SVM）,then developed a predicting model.On this basis,we also compared different models consisted of different kernel functions.Experiments demonstrated that this approach performed well on classification and prediction,which would contribute to forecast the occurrence of AHE.

关 键 词：急性低血压 平均动脉压 统计特征 支持向量机 预测 

分 类 号：R318[医药卫生—生物医学工程]