空间分位数自回归模型中的内分位点压缩技术
Interquantile Shrinkage in General Spatial Quantile Autoregressive Regression Models
-
摘要: 为处理空间相依数据中不同分位数水平下的特定效应, 空间分位数自回归(SQAR) 模型应运而生. 传统的分位数回归模型更注重于不同分位数下回归系数的估计, 往往忽略对条件分位数回归模型的全局判断. 如果采用传统的Wald多重检验方法判断分位数模型系数的差异性, 不仅会增加计算负担, 而且会带来更高的错误发现率(FDR). 为此, 我们将参数估计和差异检验转化为惩罚问题, 基于工具变量和内分点压缩技术提出一种两阶段内分位点估计方法, 包括融合自适应性LASSO (FAL)估计器和融合自适应性最大范数(FAS) 估计器. 该估计方法可以借助工具变量消除自回归模型带来的内生性, 并且在不同分位数水平下, 借助惩罚正则项对无显著差异的分位数参数进行适当地压缩合并.可以说, 该方法能够在判断内分位点是否存在显著差异的同时完成参数估计, 使得整个估计过程高效完成. 本文给出FAL和FAS估计量的Oracle性质, 并通过大量的蒙特卡洛模拟试验以及对犯罪数据集的分析验证了两阶段内分位点估计方法的有效性.Abstract: To address variable dependence and evaluate quantile-specific effects by covariates, spatial quantile autoregressive (SQAR) models have been introduced. Conventional quantile regression focuses solely on the fitting models but ignores the examinations of multiple conditional quantile functions, which provides a comprehensive view of the relationship between the response and covariates. Thus, it is necessary to investigate the different regression slopes at different quantiles, especially in situations where the quantile coeffcients share some common feature. However, traditional Wald multiple tests not only increase the burden of computation but also leads to a higher False Discovery Rate (FDR). In this paper, we transform the estimation and examination problem into a penalization problem, which estimates the parameters at different quantiles and identifies the interquantile commonality at the same time. Based on instrumental variables and shrinkage techniques, we propose a Two-stage Interquantile Estimation (TSIE) method, including Fused Adaptive LASSO (FAL) and Fused Adaptive Sup-norm (FAS) estimators. The oracle properties of the TSIE method were established. Through Monte Carlo simulations and numerical investigations of Crime data, it is demonstrated that the proposed method leads to higher estimation effciency than the traditional quantile regression.