基于惩罚回归的老年抑郁症影响因素变量选择  

Variable Selection of Factors Influencing Geriatric Depression Based on Penalized Regression

作　　者：伍晓婷 杨宜平[1,2] WU Xiaoting;YANG Yiping(School of Mathematics and Statistics,Chongqing Technology and Business University,Chongqing 400067,China;Chongqing Key Laboratory of Statistical Intelligent Computing and Monitoring,Chongqing 400067,China)

机构地区：[1]重庆工商大学数学与统计学院,重庆400067 [2]统计智能计算与监测重庆实验室,重庆400067

出　　处：《许昌学院学报》2025年第2期16-20,共5页Journal of Xuchang University

基　　金：重庆市自然科学基金项目(cstc2021jcyj-msxmX0079)。

摘　　要：针对老年人是抑郁症的高发人群问题,使用2011-2014年美国国家健康和营养检查调查数据,选取指标并建立惩罚回归模型研究影响老年人抑郁的影响因素,使用SCAD、MCP、Lasso、弹性网及自适应Lasso五种惩罚变量选择方法对老年群体抑郁进行分析,得到模型的估计结果和错误率.研究表明性别、体重和睡眠障碍情况等变量对老年抑郁研究具有重要影响,其中性别对老年抑郁的影响最为显著.通过比较错误率,发现弹性网的效果最佳,其次是自适应Las-so.最后,根据变量选择结果提出了应对和预防老年抑郁的相关建议.Targeting the high prevalence of depression among the elderly,cross-sectional data is employed from the 2011-2014 National Health and Nutrition Examination Survey(NHANES)to identify pertinent indicators and a penalty regression model is established for investigating the influencing factors of depression in this demographic.Five penalty variable selection methods,SCAD,MPC,Lasso,Elastic Net,and Adaptive Lasso,are implemented to analyze depression in the elderly,leading to the acquisition of the models’estimated results.Subsequently,the estimated error rates for each variable selection method are calculated.The findings reveal that gender,weight,sleep disorders,and diabetes exert significant effects on the study of depression among the elderly,with gender emerging as the most prominent factor.A comparative analysis of the error rates of the five variable selection methods indicates that Elastic Net achieved the best performance,closely followed by Adaptive Lasso,while the other methods exhibited minimal differences in effectiveness.Based on the outcomes of the variable selection,pertinent recommendations are offered for addressing and preventing depression among the elderly.

关 键 词：弹性网 变量选择 自适应Lasso SCAD惩罚 老年抑郁 

分 类 号：O212.7[理学—概率论与数理统计]