可解释机器学习及其在脑血管疾病预后预测中应用  

Interpretable machine learning and its application in prognosis prediction of cerebrovascular diseases

作　　者：罗枭 朱荣慧 台耀永 吴骋[1] 贺佳[1] LUO Xiao;ZHU Ronghui;TAI Yaoyong;WU Cheng;HE Jia(Military Health Statistics Teaching and Research Office,Department of Health Service,Naval Medical University,Shanghai 200433,China)

机构地区：[1]海军军医大学卫勤系军队卫生统计学教研室,上海200433

出　　处：《中国数字医学》2024年第12期80-86,共7页China Digital Medicine

基　　金：国家自然科学基金(82373687);上海市产业协同创新项目(2021-cyxt1-kj10)。

摘　　要：长期以来,具有“黑箱”性质的复杂机器学习模型在医疗健康领域蓬勃应用的同时也引发了广泛的关注和担忧。为解决这一问题,可解释机器学习相关研究发展迅速。对于应用复杂机器学习模型的脑血管疾病领域临床医生来说,了解可解释机器学习的基本概念、方法,以及批判性的认识至关重要。为此,本文综述了可解释机器学习的核心概念、方法及其在脑血管疾病中的应用现况与挑战,以期为可解释机器学习在该领域的应用与发展提供参考。Over the years,the widespread application of complex machine learning models with "black box" properties in medical and healthcare field has raised extensive attention and concern.In response to this issue,there has been rapid development in the field of interpretable machine learning.For clinicians in the field of cerebrovascular diseases who apply complex machine learning models,understanding the basic concepts,methods,and critical awareness of interpretable machine learning are crucial.Therefore,this article provides an overview of the core concepts and methods of interpretable machine learning,as well as its current applications and challenges in the field of cerebrovascular diseases,in order to provide insights for the application and development of interpretable machine learning in this area.

关 键 词：可解释机器学习 脑血管疾病 临床预测模型 

分 类 号：R197.3[医药卫生—卫生事业管理] R319[医药卫生—公共卫生与预防医学]