联合元数据与体检数据的支气管哮喘患病预测  

Prediction of Bronchial Asthma by Combining Metadata and Physical Examination Data

作　　者：涂所成 唐洁 Tu Suocheng;Tang Jie(Library and Big Data Center,Jiangxi Modern Polytechnic College,Nacchang,Jiangxi 330095,China;College of Modern Information Industry,Guangzhou College of Commerce)

机构地区：[1]江西现代职业技术学院图书与大数据中心,江西南昌330095 [2]广州商学院现代信息产业学院

出　　处：《计算机时代》2025年第4期56-58,63,共4页Computer Era

基　　金：广州商学院2023年度校级科研项目(2023XJYB30)。

摘　　要：针对支气管激发试验耗时长、临床开展率低的问题,提出一种结合元数据(性别、年龄、身高、体重等)和体检数据的支气管哮喘预测模型。首先采用Pearson、Spearman、Kendall相关分析方法解析元数据与哮喘的相关性;接着利用随机森林评估体检数据中各功能参数对哮喘预测的影响;再通过支持向量机算法筛选可改善模型的特征亚组;最终联合元数据和体检数据,构建神经网络预测模型,测试精度达80%左右。该模型有望为支气管哮喘的临床诊断提供有效参考。This article proposes a bronchial asthma prediction model that combines metadata(gender,age,height,weight,etc.)and physical examination data to address the issues of long duration and low clinical application rate of bronchial provocation tests.Firstly,the correlation between metadata and asthma was analyzed using Pearson,Spearman,and Kendall correlation analysis methods;Then use random forest algorithm to evaluate the impact of various functional parameters in physical examination data on asthma prediction;Then use support vector machine algorithm to find feature subgroups that can improve the model.Finally,joint metadata and physical examination data modeling,using neural networks as the prediction model,achieved a test accuracy of about 80%.This model is expected to provide effective reference for clinical diagnosis of bronchial asthma.

关 键 词：激发试验 随机森林 支持向量机 神经网络 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]