基于Stacking融合方法的幸福感预测与提升路径研究

Happiness Prediction and Promotion Path Based on Stacking Fusion Method

幸福感作为反映居民生活水平的一个重要指标,对社会治理和社会发展具有一定的指导作用。新时代背景下,我国社会主要矛盾已经发生变化,如何更好地满足人民美好生活需要,提升居民幸福感成为政府关注的重点。本文基于中国综合社会调查(CGSS) 2015数据,通过嵌入法进行特征选择,使用逻辑回归、支持向量机、随机森林、LightGBM以及Stacking融合方法进行综合评估预测。在此基础上发现与幸福感关联度最高的三个特征并以此研究了幸福感的提升路径,对提升国民幸福感有一定参考指导意义。

As an important indicator of residents’ living standards, happiness plays a guiding role in social governance and social development. Under the background of the new era, the main social con-tradiction in China has changed. How to better meet the people’s needs for a better life and improve residents’ happiness has become the focus of the government. Based on the 2015 data from China Comprehensive Social Survey (CGSS), this paper conducts feature selection by embedding method and conducts comprehensive assessment and prediction by using logistic regression, support vector machine, random forest, LightGBM and stacking fusion method. On this basis, it finds the three features that are most closely related to happiness and studies the path to enhance happiness, which has certain reference significance for improving national happiness.