遗传 ›› 2018, Vol. 40 ›› Issue (9): 693-703.doi: 10.16288/j.yczz.18-139

• 综述 •    下一篇

组学时代下机器学习方法在临床决策支持中的应用

赵学彤1,2(),杨亚东1,2,渠鸿竹1,2,方向东1,2()   

  1. 1. 中国科学院北京基因组研究所,中国科学院基因组科学与信息重点实验室,北京 100101
    2. 中国科学院大学,北京 100049
  • 收稿日期:2018-05-17 修回日期:2018-07-23 出版日期:2018-09-20 发布日期:2018-07-30
  • 作者简介:赵学彤,博士研究生,专业方向:复杂疾病多组学数据整合与解析。E-mail: zhaoxuetong@big.ac.cn
  • 基金资助:
    国家“精准医学研究”重点研发计划项目(2016YFC0901700);国家“精准医学研究”重点研发计划项目(2016YFC0901603);国家“精准医学研究”重点研发计划项目(2017YFC0907502);国家“精准医学研究”重点研发计划项目(2017YFC0908402);国家“精准医学研究”重点研发计划项目(2017YFC0907405)

Applications of machine learning in clinical decision support in the omic era

Zhao Xuetong1,2(),Yang Yadong1,2,Qu Hongzhu1,2,Fang Xiangdong1,2()   

  1. 1. CAS Key Laboratory of Genome Sciences and Information, Beijing Institute of Genomics, Chinese Academy of Sciences, Beijing 100101, China
    2. University of Chinese Academy of Sciences, Beijing 100049, China
  • Received:2018-05-17 Revised:2018-07-23 Online:2018-09-20 Published:2018-07-30
  • Supported by:
    Supported by the National Key Research and Development Program of China(2016YFC0901700);Supported by the National Key Research and Development Program of China(2016YFC0901603);Supported by the National Key Research and Development Program of China(2017YFC0907502);Supported by the National Key Research and Development Program of China(2017YFC0908402);Supported by the National Key Research and Development Program of China(2017YFC0907405)

摘要:

随着组学技术的不断发展,对于不同层次和类型的生物数据的获取方法日益成熟。在疾病诊治过程中会产生大量数据,通过机器学习等人工智能方法解析复杂、多维、多尺度的疾病大数据,构建临床决策支持工具,辅助医生寻找快速且有效的疾病诊疗方案是非常必要的。在此过程中,机器学习等人工智能方法的选择显得尤为重要。基于此,本文首先从类型和算法角度对临床决策支持领域中常用的机器学习等方法进行简要综述,分别介绍了支持向量机、逻辑回归、聚类算法、Bagging、随机森林和深度学习,对机器学习等方法在临床决策支持中的应用做了相应总结和分类,并对它们的优势和不足分别进行讨论和阐述,为临床决策支持中机器学习等人工智能方法的选择提供有效参考。

关键词: 疾病, 机器学习, 人工智能, 临床决策支持

Abstract:

With the development of the omic technologies, the acquisition approaches of various biological data on different levels and types are becoming more mature. As a large amount of data will be produced in the process of diagnosis and treatment of diseases, it is necessary to utilize the artificial intelligence such as machine learning to analyze complex, multi-dimensional and multi-scale data and to construct clinical decision support tools. It will provide a method to figure out rapid and effective programs in diagnosis and treatment. In this process, the choice of artificial intelligence seems to be particularly important, such as machine learning. The article reviews the type and algorithm of machine learning used in clinical decision support, such as support vector machines, logistic regression, clustering algorithms, Bagging, random forests and deep learning. The application of machine learning and other methods in clinical decision support has been summarized and classified. The advantages and disadvantages of machine learning are elaborated. It will provide a reference for the selection between machine learning and other artificial intelligence methods in clinical decision support.

Key words: diseases, machine learning, artificial intelligence, clinical decision support