| 基于机器学习的宁夏地区老年人生物学年龄研究 |
| |
| 引用本文： | 张俭1,张婕2,申勐韬3,李小龙4,赵瑜4,李锋2. 基于机器学习的宁夏地区老年人生物学年龄研究[J]. 现代预防医学, 2023, 0(1): 6-9. DOI: 10.20043/j.cnki.MPM.202206210 |
| |
| 作者姓名： | 张俭1 张婕2 申勐韬3 李小龙4 赵瑜4 李锋2 |
| |
| 作者单位： | 1.宁夏医科大学第二临床医学院,宁夏 银川 750010;2．宁夏医科大学总医院医学实验中心;3．宁夏银川西夏区正茂社区卫生服务站;4．宁夏医科大学公共卫生与管理学院 |
| |
| 基金项目： | 国家重点研发计划(2018YFC2000200); |
| |
| 摘 要： | 目的 生物学年龄(Biological age, BA)可以更有效的判断个体真正的衰老状态,精准预测BA有助于为老年个体早期制定有针对性的预防措施,目前关于老年人分亚群对生物学年龄与生化指标相关性研究较少。利用机器学习算法计算宁夏地区老年人的生物学年龄,并识别相关生物化学指标分亚群进行分析。方法 纳入2020年宁夏地区老年人健康体检者共4 060名作为研究对象,采集空腹静脉血、尿液检测生物化学指标,利用随机森林(Random Forest, RF)算法筛选与BA相关的生物学指标,计算生物学年龄,并对RF算法的预测精度进行评估。结果 在老年人的不同亚群(低龄、中龄、高龄)中,各年龄段生物学指标在生物学年龄的重要性各有不。研究采用平均绝对误差(Mean Absolute Error,MAE)、均方根误差(Root Mean Square Error,RMSE)、相关系数(Coefficient of association,R2)进行模型的效能评估。结论 利用随机森林计算老年人生物学年龄并分析相关指标,可以更加精准定位老年人中高危人群及对健康老龄化有一定的指导。
|
| 关 键 词： | 机器学习 随机森林 生物学年龄 重要性 |
Machine learning-based research on the biological age of elderly people,Ningxia |
| |
| ZHANG Jian,ZHANG Jie,SHEN Meng-tao,LI Xiao-long,ZHAO Yu,LI Feng. Machine learning-based research on the biological age of elderly people,Ningxia[J]. Modern Preventive Medicine, 2023, 0(1): 6-9. DOI: 10.20043/j.cnki.MPM.202206210 |
| |
| Authors: | ZHANG Jian ZHANG Jie SHEN Meng-tao LI Xiao-long ZHAO Yu LI Feng |
| |
| Affiliation: | *Second Clinical Medical School of Ningxia Medical University, Yinchuan, Ningxia 750010, China |
| |
| Abstract: | Objective To calculate the biological age of older adults in Ningxia by machine learning algorithms and to identify relevant biochemical indicators for analysis, on the basis that biological age (BA) can determine the true aging status of individuals and an accurate prediction of BA can help develop targeted preventive measures for older individuals at an early stage, while for the elderly, there are limited investigations on the association between biological age and biochemical indicators. Methods Fasting venous blood and urine was obtained to evaluate biochemical indicators, and other things in a total of 4060 elderly health checks in the Ningxia area in 2020. The BA-related datasets were screened using Random Forest (RF) methods. Results Importance of the biological index was different in classification population of the elderly (young-old, middle-old, the oldest old). In this study, it was used to evaluate the effectiveness of the model by Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of association (R2). Conclusion Utilizing random forest to compute the biological age of the elderly and assess the relevant indicators can be more precise in identifying high-risk groups among the elderly and providing some advice for healthy aging. |
| |
| Keywords: | Machine learning Random Forest Biological age Importance |
|
| 点击此处可从《现代预防医学》浏览原始摘要信息 |
|
点击此处可从《现代预防医学》下载免费的PDF全文 |
|
