arrow
返回

Benchmarking Machine Learning Models for Obesity Classification with SHAP-Based Interpretability

delete2025-12-27
delete0
PRE
AI
V
Vaibhav C. Gandhi
Y
Yogesh Chaudhari
K
Kumar, Ajay *
H
Hitarth Revakar
A
Ankit D. Oza
S
Saneh Lata Yadav
DOI:10.1007/s44196-025-01078-xdelete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
肥胖是一种全球性流行病,对普通人群在慢性代谢疾病和低长寿方面构成严重威胁。临床干预和预防性医疗保健需要早期识别和适当的风险优先排序。本文比较了六种机器学习模型,即逻辑回归、决策树、随机森林、支持向量机、K近邻和梯度提升,使用了来自墨西哥、秘鲁和哥伦比亚的公开肥胖数据集(n = 2111,17个属性)。为克服某些类别占主导地位的问题,在预处理阶段使用了合成少数类过采样技术(SMOTE),以使模型得到公平考虑。梯度提升模型表现最佳,在所有模型中准确率达到95.93%,精确率为0.96,召回率为0.96,F1得分为0.96。SHAP(SHapley Additive Explanations)分析用于提高可解释性,每个预测因子表明最重要的决定因素;频繁高热量食物摄入(FAVC, |human|)(0.18)、身体活动频率(FAF, 0.15)、超重家族史(0.12)和水分摄入水平(0.09)具有显著意义。这些定量见解不仅增强了可解释性,也符合临床关于行为和遗传风险因素的知识。所提出的可解释框架可以为设计数据驱动的决策支持工具提供坚实基础和清晰背景,这些工具可用于预防肥胖、制定个性化咨询以及设计满足人群需求的政策。
Keyword:
Obesity prediction
Machine learning benchmarking
Gradient boosting (GB)
Explainable AI (XAI)
SHAP analysis
Clinical decision support

期刊

International Journal of Computational Intelligence Systems 封面图
International Journal of Computational Intelligence Systems
IF:
3
论文数:
362
被引数:
2.7K

机构

C
chandigarh university
学者数:
402
论文数: 387
被引数: 0
M
Manipal University Jaipur
学者数:
2.3K
论文数: 1.7K
被引数: 1.1K
引用论文

引用论文

A Systematic Review on Machine Learning Intelligent Systems for Heart Disease Diagnosis用于心脏病诊断的机器学习智能系统的系统综述
err2025-03-01
err0
PREAI
errSharma, Abhinav; Dhanka, Sanjay; Kumar, Ankur; Nain, Monika; Dhanka, Balan; Bhardwaj, Vibhor Kumar; Maini, Surita; Arora, Ajat Shatru
err分享
err收藏
Sepsis and obesity: a scoping review of diet-induced obesity murine models
err2024-02-23
err0
errOAAI
errMikaela Eng; Keshikaa Suthaaharan; Logan Newton; Fatima Sheikh; Alison Fox-Robichaud
err分享
err收藏
Machine learning applications for COVID-19 outbreak management
err2022-06-10
err0
errOAAI
errArash Heidari; Nima Jafari Navimipour; Mehmet Unal; Shiva Toumaj
err分享
err收藏
Obesity and Leukemia: Biological Mechanisms, Perspectives, and Challenges
err2023-12-30
err8
errOAAI
errTsilingiris, Dimitrios; Vallianou, Natalia G.; Spyrou, Nikolaos; Kounatidis, Dimitris; Christodoulatos, Gerasimos Socrates; Karampela, Irene; Dalamaga, Maria
err分享
err收藏
err分享
err收藏
Machine Learning Techniques for Prediction of Early Childhood Obesity
err2017-12-19
err0
errOAAI
errS. Mukhopadhyay; A. Carroll; S. Downs; T. M. Dugan
err分享
err收藏
Predicting Childhood Obesity Using Machine Learning: Practical Considerations
err2022-03-08
err0
errOAAI
errErika R. Cheng; Rai Steinhardt; Zina Ben Miled
err分享
err收藏
学者 查看更多内容