文章摘要
李翰芳,罗幼喜,田茂再.面板数据的贝叶斯Lasso分位回归方法[J].数量经济技术经济研究,2013,30(2):138-149
面板数据的贝叶斯Lasso分位回归方法
Bayesian Lasso Quantile Regression for Panel Data Models
  
DOI:
中文关键词: 面板数据  贝叶斯Lasso  分位回归  切片Gibbs抽样
英文关键词: Panel Data  Bayesian Lasso  Quantile Regression  Slice Gibbs Sampler
基金项目:
作者单位
李翰芳 湖北工业大学理学院 
罗幼喜 湖北工业大学理学院 
田茂再 中国人民大学统计学院 
中文摘要:
      文章讨论了含有随机效应的面板数据模型,通过引入条件Laplace先验,文章构造了一种新的贝叶斯Lasso分位回归法。与一般贝叶斯分位回归法不同的是,该方法能够更大程度的将模型中非重要解释变量系数压缩至0,从而在估计系数的同时也起到了变量选择的作用。利用积分恒等式,文章构造了一种易于实施的参数估计的切片Gibbs抽样算法。模拟结果显示,在模型含有较多变量时,新方法排除“噪声”变量的能力明显高于现有文献中其他方法。文章最后对我国各地区多个宏观经济指标的面板数据进行了建模分析,演示了新方法估计参数与挑选变量的能力。
英文摘要:
      The paper discusses the random effects panel data model and establishes a new Bayesian Lasso quantile regression method by using of conditional Laplace prior. It is different with the general Bayesian quantile regression method is that it can shrink the coefficients of non-important explanatory variables to zero more powerful。thus doing variable selection simultaneously. By using of integral identities。the paper constructs an easy slice Gibbs sampling algorithm to estimate parameters. Monte Carlo simulation study indicates that the proposed method is obviously superior to other methods in literatures when excluding “noise” variables from a lot of explanatory variables. Finally。we studies the panel data including several macroeconomic indicators of our country and demonstrates the new method’s capability of estimating parameters and doing variable selection.
查看全文