Question

问题很简单：如何从lqmm包中的汇总函数获得每个tau级别的每个变量的标准误差，下界和上界值？

              Value      Std. Error lower bound upper bound  Pr(>|t|)    
(Intercept)   value1         13.9770730  0.3399716  13.2938744     14.6603 < 2.2e-16 ***
factor(a)2    value2         -0.6249463  0.0359903  -0.6972716     -0.5526 < 2.2e-16 ***
factor(b)2    value3          0.3511184  0.0500702   0.2504986      0.4517 6.344e-09 ***

我想得到13.977的价值。无论如何可能吗？如果我使用QR2013SKDHRP $ tau，则只能获得第一列系数和其他几个值，但不能获得标准误差，下限和上限。

Answer 1

sameDomain函数在每个请求的tqu之前重新打开一个带有值矩阵的列表。根据帮助页面，此结果位于名为origin的叶子中。使用帮助页面上的示例作为起点：

summary.lqmm

评论中报告的问题似乎与"tTable"与> res <- summary(fitOi.lqmm) Warning message: In errorHandling(OPTIMIZATION$low_loop, "low", control$LP_max_iter, : Lower loop did not converge in: lqmm. Try increasing max number of iterations (500) or tolerance (1e-05) > res$tTable $`0.1` Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 16.7336088 0.7300274 15.2665637 18.2006538 7.620900e-28 age 0.5221987 0.0783630 0.3647224 0.6796751 2.201244e-08 $`0.5` Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 16.8119676 0.72758041 15.3498401 18.2740952 5.312804e-28 age 0.6188025 0.08742306 0.4431193 0.7944857 5.018630e-09 $`0.9` Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 16.8267888 0.7346688 15.3504165 18.3031611 7.894189e-28 age 0.7961899 0.1008876 0.5934487 0.9989311 2.796707e-10的选择无关。您会在lqmm帮助页面上查看应用于该示例的lqm输出，获得相同类型的结果：

summary.lqm

演示如何访问输出中第一个矩阵的Std.Error列：

?lqm

如果您想在一个答案中获得fit.lqm #------------------- Call: lqm(formula = y ~ x, data = test, tau = p, control = list(verbose = FALSE, loop_tol_ll = 1e-09), fit = TRUE) Fixed effects: tau = 0.25 tau = 0.50 tau = 0.75 (Intercept) 29.322072 29.954761 30.628379 x 1.124451 1.182257 1.251657 Degrees of freedom: 500 total; 498 residual #------------ names(summary(fit.lqm)) #-------------- [1] "0.25" "0.50" "0.75" "theta" "scale" "call" [7] "term.labels" "terms" "nobs" "edf" "dim_theta" "rdf" [13] "tau" "x" "y" "weights" "levels" "InitialPar" [19] "control" "tTable" summary(fit.lqm)$tTable #---------------- [[1]] Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 29.322072 0.1071627 29.1067201 29.537423 1.089432e-79 x 1.124451 0.2137049 0.6949946 1.553907 3.134769e-06 [[2]] Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 29.954761 0.1001000 29.7536022 30.155919 1.358822e-81 x 1.182257 0.1447352 0.8914013 1.473114 1.057212e-10 [[3]] Value Std. Error lower bound upper bound Pr(>|t|) (Intercept) 30.628379 0.0833481 30.4608850 30.795874 5.813793e-86 x 1.251657 0.1736253 0.9027443 1.600571 3.150189e-09的第一个值的“标准错误值，下限和上限”。它只会是：

> summary(fit.lqm)$tTable[[1]][, "Std. Error"]
(Intercept)           x 
  0.1296308   0.2851553

lqmm包中摘要的标准错误值

1 个答案: