Comprehensive analysis of feature-algorithm interactions for fall detection across age groups via machine learning

IF 1.8 4区 计算机科学 Q3 COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE Computational Intelligence Pub Date : 2024-10-23 DOI:10.1111/coin.12697
Erhan Kavuncuoğlu
{"title":"Comprehensive analysis of feature-algorithm interactions for fall detection across age groups via machine learning","authors":"Erhan Kavuncuoğlu","doi":"10.1111/coin.12697","DOIUrl":null,"url":null,"abstract":"<p>Fall detection in daily activities hinges on both feature selection and algorithm choice. This study delves into their intricate interplay using the Sisfall dataset, testing 10 machine learning algorithms on 26 features encompassing diverse falls and age groups. Individual feature analysis yields key insights. RFC with the autocorrelation feature outperformed the other classifiers, achieving 97.94% accuracy and 97.51% sensitivity (surpassing F3-SVM at 96.18% and F17-LightGBM at 95.79%). The F3-SVM exhibited exceptional specificity (98.72%) for distinguishing daily activities. Time-series features employed by SVM achieved a peak accuracy of 98.60% on unseen data, exceeding motion, basic statistical, and frequency domain features. Feature combinations further excel: the Quintuple approach, fusing top-performing features, reaches 98.69% accuracy, 98.28% sensitivity, and 99.08% specificity with the ETC, demonstrating notable sensitivity owing to its adaptability. This study underscores the crucial interplay of features and algorithms, with the Quintuple-ETC approach emerging as the most effective. Rigorous hyperparameter tuning strengthens its performance in real-world fall-detection applications. Furthermore, the study investigates algorithm transferability, training models on young participants' data and applying them to the elderly—a significant challenge in machine learning. This highlights the importance of understanding the data transfer between age groups in healthcare, aging management, and medical diagnostics.</p>","PeriodicalId":55228,"journal":{"name":"Computational Intelligence","volume":"40 5","pages":""},"PeriodicalIF":1.8000,"publicationDate":"2024-10-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computational Intelligence","FirstCategoryId":"94","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1111/coin.12697","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE","Score":null,"Total":0}
引用次数: 0

Abstract

Fall detection in daily activities hinges on both feature selection and algorithm choice. This study delves into their intricate interplay using the Sisfall dataset, testing 10 machine learning algorithms on 26 features encompassing diverse falls and age groups. Individual feature analysis yields key insights. RFC with the autocorrelation feature outperformed the other classifiers, achieving 97.94% accuracy and 97.51% sensitivity (surpassing F3-SVM at 96.18% and F17-LightGBM at 95.79%). The F3-SVM exhibited exceptional specificity (98.72%) for distinguishing daily activities. Time-series features employed by SVM achieved a peak accuracy of 98.60% on unseen data, exceeding motion, basic statistical, and frequency domain features. Feature combinations further excel: the Quintuple approach, fusing top-performing features, reaches 98.69% accuracy, 98.28% sensitivity, and 99.08% specificity with the ETC, demonstrating notable sensitivity owing to its adaptability. This study underscores the crucial interplay of features and algorithms, with the Quintuple-ETC approach emerging as the most effective. Rigorous hyperparameter tuning strengthens its performance in real-world fall-detection applications. Furthermore, the study investigates algorithm transferability, training models on young participants' data and applying them to the elderly—a significant challenge in machine learning. This highlights the importance of understanding the data transfer between age groups in healthcare, aging management, and medical diagnostics.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
通过机器学习全面分析各年龄组跌倒检测中特征与算法之间的相互作用
日常活动中的跌倒检测取决于特征选择和算法选择。本研究利用 Sisfall 数据集深入探讨了它们之间错综复杂的相互作用,在 26 个特征上测试了 10 种机器学习算法,这些特征包括不同的跌倒和年龄组。对单个特征的分析得出了关键的见解。带有自相关特征的 RFC 的表现优于其他分类器,准确率达到 97.94%,灵敏度达到 97.51%(超过了 96.18% 的 F3-SVM 和 95.79% 的 F17-LightGBM)。F3-SVM 在区分日常活动方面表现出了极高的特异性(98.72%)。SVM 采用的时间序列特征在未见数据上达到了 98.60% 的峰值准确率,超过了运动、基本统计和频域特征。特征组合的效果更加突出:融合了最佳特征的 Quintuple 方法与 ETC 的准确率达到了 98.69%,灵敏度达到了 98.28%,特异性达到了 99.08%,由于其适应性强,灵敏度显著提高。这项研究强调了特征和算法之间的重要相互作用,其中五元-ETC 方法最为有效。严格的超参数调整增强了其在实际跌倒检测应用中的性能。此外,该研究还调查了算法的可移植性,即在年轻参与者的数据上训练模型,然后将其应用于老年人--这在机器学习中是一项重大挑战。这凸显了了解医疗保健、老龄化管理和医疗诊断中不同年龄组之间数据转移的重要性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Computational Intelligence
Computational Intelligence 工程技术-计算机:人工智能
CiteScore
6.90
自引率
3.60%
发文量
65
审稿时长
>12 weeks
期刊介绍: This leading international journal promotes and stimulates research in the field of artificial intelligence (AI). Covering a wide range of issues - from the tools and languages of AI to its philosophical implications - Computational Intelligence provides a vigorous forum for the publication of both experimental and theoretical research, as well as surveys and impact studies. The journal is designed to meet the needs of a wide range of AI workers in academic and industrial research.
期刊最新文献
Issue Information Comprehensive analysis of feature-algorithm interactions for fall detection across age groups via machine learning An Efficient and Robust 3D Medical Image Classification Approach Based on 3D CNN, Time-Distributed 2D CNN-BLSTM Models, and mRMR Feature Selection Modified local Granger causality analysis based on Peter-Clark algorithm for multivariate time series prediction on IoT data A Benchmark Proposal for Non-Generative Fair Adversarial Learning Strategies Using a Fairness-Utility Trade-off Metric
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1