人口老龄化背景下老年人幸福感影响因素探析——基于改进GRA-二分类Logistic模型与RUSBoost算法

Exploring the Factors Influencing the Well-Being of the Elderly in the Context of Population Ageing —Based on Improved GRA-Disaggregated Logistic Model with RUSBoost Algorithm

摘要: 随着中国社会老龄化程度的不断加深，老年人口总量和比例逐年增加。老有所养，事关亿万百姓福祉，也是人口老龄化社会面临的一道必答题。本文选取CHARLS 2018年全国调查的数据，运用改进GRA-Logistic回归模型和RUSBoost算法探究老年人幸福感与健康行为、家庭关系、人际交往、工作概况、医疗养老服务之间的关系，并分析比较二者的优劣。结果表明，医疗养老服务对老年人幸福感影响程度最大，人际交往最小；性别、受教育程度和婚姻等因素与老年人幸福感有关。模型分析比较显示，改进GRA-Logistic二分类模型准确度比一般Logistic回归模型高，GRA-Logistic二分类模型的分类情况比RUSBoost算法好。基于此，本文提出相关建议，为培育和健全养老制度、提高老年人幸福感提供更好的决策依据。

Abstract: With the deepening of aging in Chinese society, the total number and proportion of elderly people are increasing year by year. The well-being of hundreds of millions of people is at stake, and it is also a necessary answer to the question of population ageing society. This paper uses data from the CHARLS 2018 national survey to investigate the relationship between older people’s well- being and health behaviours, family relationships, interpersonal interactions, work profiles and medical and elderly care services using an improved GRA-Logistic regression model and the RUSBoost algorithm, and to analyze and compare the advantages and disadvantages of the two. The results show that medical and elderly care services have the greatest impact on older people’s well-being and interpersonal interactions the least; factors such as gender, education level and marriage are related to older people’s well-being. The model analysis comparison shows that the improved GRA-Logistic dichotomous classification model is more accurate than the general logistic regression model, and the GRA-Logistic dichotomous classification model has a better classification than the RUSBoost algorithm. Based on this, this paper puts forward relevant suggestions to provide a better basis for decision making to foster and improve the elderly system and enhance the well-being of the elderly.