Computer Engineering and Applications ›› 2017, Vol. 53 ›› Issue (20): 24-30.DOI: 10.3778/j.issn.1002-8331.1708-0229

Previous Articles     Next Articles

Study of computerized methods to predict in-hospital mortality in intensive care unit

XIE Junqing1,2, LIN Ke1,2, KONG Guilan1   

  1. 1.Medical Informatics Center, Peking University, Beijing 100191, China
    2.Department of Epidemiology & Biostatistics, School of Public Health, Peking University, Beijing 100191, China
  • Online:2017-10-15 Published:2017-10-31


谢俊卿1,2,蔺  轲1,2,孔桂兰1   

  1. 1.北京大学 医学信息学中心,北京 100191
    2.北京大学 公共卫生学院,北京 100191

Abstract: Through literature review, the basic concepts, procedures and common methods involved in the development of mortality prediction models are summarized. First, the sources of predictor variables and their corresponding screening approaches are discussed. Second, on the basis of the logistic regression model which is most familiar to clinicians, the basic framework and advantages and disadvantages of the neural network, decision tree and support vector machine are elaborated. Finally, the common metrics used to evaluate the performance of the model are described, as well as the way to validate the model.

Key words: mortality prediction, logistic regression, machine learning, Intensive Care Unit(ICU)

摘要: 通过文献综述,总结了在构建ICU患者住院死亡风险预测模型时所涉及的基本概念、步骤和常用方法。讨论了预测变量的来源及其筛选办法。在临床医生最为熟知的逻辑回归模型的基础上,阐述了人工神经网络、决策树和支持向量机三种机器学习模型的基本框架以及优劣势。描述了用于评价模型性能好坏的各种指标以及验证模型性能的方式。

关键词: 死亡预测, 逻辑回归, 机器学习, 重症监护室