如何使用 stargazer 创建分位数回归表？

Question

我使用 quantreg 包计算了以下分位数回归

    qr_10 = rq(inno_DELTA ~ deDomains + R_and_D_pc + Pop_dens + Agr_GDP + Manufacturing_GDP + Service_GDP + Infr_Area_Percent + Res_pc + Debt_GDP + GOV_EXP_GDP + firms_total + factor(landkreis) + factor(jahr), tau = 0.10, data = df_ip_c)
    qr_25 = rq(inno_DELTA ~ deDomains + R_and_D_pc + Pop_dens + Agr_GDP + Manufacturing_GDP + Service_GDP + Infr_Area_Percent + Res_pc + Debt_GDP + GOV_EXP_GDP + firms_total + factor(landkreis) + factor(jahr), tau = 0.25, data = df_ip_c)
    qr_50 = rq(inno_DELTA ~ deDomains + R_and_D_pc + Pop_dens + Agr_GDP + Manufacturing_GDP + Service_GDP + Infr_Area_Percent + Res_pc + Debt_GDP + GOV_EXP_GDP + firms_total + factor(landkreis) + factor(jahr), tau = 0.5, data = df_ip_c)
    qr_75 = rq(inno_DELTA ~ deDomains + R_and_D_pc + Pop_dens + Agr_GDP + Manufacturing_GDP + Service_GDP + Infr_Area_Percent + Res_pc + Debt_GDP + GOV_EXP_GDP + firms_total + factor(landkreis) + factor(jahr), tau = 0.75, data = df_ip_c)
    qr_95 = rq(inno_DELTA ~ deDomains + R_and_D_pc + Pop_dens + Agr_GDP + Manufacturing_GDP + Service_GDP + Infr_Area_Percent + Res_pc + Debt_GDP + GOV_EXP_GDP + firms_total + factor(landkreis) + factor(jahr), tau = 0.95, data = df_ip_c)

我正在尝试使用stargazer 显示这些回归。我尝试运行的代码如下：

stargazer(qr_10,qr_25,qr_50,qr_75,qr_95, rq.se = "iid", type = "text", title="Regression Results" ,initial.zero = F,single.row=TRUE, out="table_quantile_regression.html")

但是，我收到以下错误消息

Error in base::backsolve(r, x, k = k, upper.tri = upper.tri, transpose = transpose,  : 
  singular matrix in 'backsolve'. First zero in diagonal [421]

我假设此错误消息与stargazer 中的标准错误函数rq.se 有关，例如summary(qr_10, se = "iid") 工作正常。

有人有解决这个问题的方法吗？ 谢谢你。

Answer 1

我想这是由于您的数据。 “奇异矩阵”表明回归存在一些问题。最小的例子工作正常，stargazer 支持quantreg：

library(quantreg)
library(stargazer)

stargazer(mtcars, type="text")

qr_10 = rq(mpg ~ factor(cyl) + disp, tau = 0.10, data = mtcars)
qr_25 = rq(mpg ~ factor(cyl) + disp, tau = 0.25, data = mtcars)
qr_50 = rq(mpg ~ factor(cyl) + disp, tau = 0.50, data = mtcars)
qr_75 = rq(mpg ~ factor(cyl) + disp, tau = 0.75, data = mtcars)
qr_95 = rq(mpg ~ factor(cyl) + disp, tau = 0.95, data = mtcars)

stargazer(qr_10,qr_25,qr_50,qr_75,qr_95, rq.se = "iid", type = "text", title="Regression Results", initial.zero = F,single.row=TRUE)
stargazer(qr_10,qr_25,qr_50,qr_75,qr_95, type = "text", title="Regression Results", initial.zero = F,single.row=TRUE)

Regression Results
=======================================================================================================
                                                Dependent variable:                                    
             ------------------------------------------------------------------------------------------
                                                        mpg                                            
                   (1)               (2)               (3)               (4)                (5)        
-------------------------------------------------------------------------------------------------------
factor(cyl)6 -2.462** (.992)   -2.105 (1.324)   -4.365*** (1.453) -8.598*** (1.605)  -14.011*** (1.024)
factor(cyl)8  -2.243 (1.735)   -2.608 (2.317)    -3.885 (2.542)   -11.437*** (2.809) -18.811*** (1.792)
disp         -.026*** (.006)   -.027*** (.009)   -.033*** (.009)     -.012 (.010)       .013* (.007)   
Constant     24.630*** (.858) 25.705*** (1.145) 29.916*** (1.256) 31.575*** (1.388)   33.011*** (.885) 
-------------------------------------------------------------------------------------------------------
Observations        32               32                32                 32                 32        
=======================================================================================================

您能否为您的数据提供更多信息或汇总统计信息？

查看https://stats.stackexchange.com/questions/78022/cause-of-singularity-in-matrix-for-quantile-regression 和相关帖子。最好的问候