基于机器学习的决策树算法在血流感染预后预测中应用现状及展望  

Current situation and prospect of application of decision tree algorithm based on machine learning in prognosis prediction of bloodstream infection

作　　者：范帅华 郭伟[1] 郭军[2] Shuaihua Fan;Wei Guo;Jun Guo(School of Clinical Medicine,Tsinghua University,Beijing 100084,China;Department of Respiratory and Critical Care Medicine,Beijing Tsinghua Changgung Hospital Affiliated with Tsinghua University,Beijing 102218,China;Department of Department of Geriatric Medicine,Beijing Tsinghua Changgung Hospital Affiliated with Tsinghua University,Beijing 102218,China)

机构地区：[1]清华大学临床医学院,清华大学附属北京清华长庚医院呼吸与危重症医学科,北京100089 [2]清华大学附属北京清华长庚医院老年医学科,清华大学临床医学院,北京102218

出　　处：《中华实验和临床感染病杂志（电子版）》2023年第5期289-293,共5页Chinese Journal of Experimental and Clinical Infectious Diseases(Electronic Edition)

基　　金：北京市卫生健康委员会高层次公共卫生技术人才建设项目培养计划(No.学科带头人-02-06)。

摘　　要：血流感染作为一种严重的全身感染,近年来其患病率逐步升高,是造成患者不良预后的主要成因之一,因此早期识别不良预后的高危患者并及时进行干预尤为重要。但对血流感染预后预测的传统统计学分析在信度及效度上无法满足临床需求,鉴于机器学习算法已在一些临床疾病的预测模型构建中取得了良好的应用效果,展示了其提升临床诊疗精准性的应用前景,本文主要针对其中的决策树算法在血流感染预后预测方面的应用现状进行综述,通过比较其与传统方法的优缺点,对其在血流感染预后预测方面的应用前景进行展望,旨在探索更好的预测方式,以便于临床早期识别高危患者,最大程度降低血流感染的病死率。As a serious systemic infection,the prevalence of bloodstream infection has gradually increased in recent years,which is one of the main causes of poor prognosis of patients,so it is particularly important to identify high-risk patients with poor prognosis early and timely.However,the traditional statistical analysis of bloodstream infection prognosis prediction can not meet the clinical needs in terms of reliability and validity,and since machine learning algorithms have achieved good application results in the construction of prediction models for some clinical problems,showing their application prospects to improve the accuracy of clinical diagnosis and treatment,this paper mainly reviews the application status of decision tree algorithm in the prognosis prediction of bloodstream infection,and prospects its application in the prediction of bloodstream infection prognosis by comparing its advantages and disadvantages with traditional methods.This review aims to explore better predictive methods for early clinical identification of high-risk patients and minimize the mortality rate of bloodstream infections.

关 键 词：人工智能 机器学习 决策树 血流感染 预后预测 

分 类 号：R446.5[医药卫生—诊断学] TP181[医药卫生—临床医学]