当前位置: X-MOL 学术J. Appl. Stat. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Bayesian variable selection in quantile regression with random effects: an application to Municipal Human Development Index
Journal of Applied Statistics ( IF 1.5 ) Pub Date : 2021-07-11 , DOI: 10.1080/02664763.2021.1950654
Marcus G L Nascimento 1 , Kelly C M Gonçalves 1
Affiliation  

According to the Atlas of Human Development in Brazil, the income dimension of Municipal Human Development Index (MHDI-I) is an indicator that shows the population's ability in a municipality to ensure a minimum standard of living to provide their basic needs, such as water, food and shelter. In public policy, one of the research objectives is to identify social and economic variables that are associated with this index. Due to the income inequality, evaluate these associations in quantiles, instead of the mean, could be more interest. Thus, in this paper, we develop a Bayesian variable selection in quantile regression models with hierarchical random effects. In particular, we assume a likelihood function based on the Generalized Asymmetric Laplace distribution, and a spike-and-slab prior is used to perform variable selection. The Generalized Asymmetric Laplace distribution is a more general alternative than the Asymmetric Laplace one, which is a common approach used in quantile regression under the Bayesian paradigm. The performance of the proposed method is evaluated via a comprehensive simulation study, and it is applied to the MHDI-I from municipalities located in the state of Rio de Janeiro.



中文翻译:

随机效应分位数回归中的贝叶斯变量选择:在市政人类发展指数中的应用

根据巴西人类发展地图集,城市人类发展指数 (MHDI-I) 的收入维度是一个指标,显示城市人口确保最低生活标准以满足其基本需求(例如水)的能力,食物和住所。在公共政策中,研究目标之一是确定与该指数相关的社会和经济变量。由于收入不平等,用分位数而不是平均值来评估这些关联可能更有趣。因此,在本文中,我们在具有分层随机效应的分位数回归模型中开发了贝叶斯变量选择。特别是,我们假设一个基于广义非对称拉普拉斯分布的似然函数,并使用尖峰和平板先验来执行变量选择。广义非对称拉普拉斯分布是比非对称拉普拉斯分布更通用的替代方法,非对称拉普拉斯分布是贝叶斯范式下分位数回归中常用的方法。通过综合模拟研究评估了所提出方法的性能,并将其应用于位于里约热内卢州的城市的 MHDI-I。

更新日期:2021-07-11
down
wechat
bug