从 R 中的黄土回归结果中提取残差标准误差

Question

我正在尝试从黄土回归模型的输出摘要中提取残差标准误差。

> summary(fit.loess[[i]])
Call:
loess(formula = dfcpm[, ncol(dfcpm)] ~ dfcpm[, i], data = dfcpm, 
    span = 0.5, degree = 1, normalize = FALSE, family = "gaussian")

Number of Observations: 88 
Equivalent Number of Parameters: 4.7 
Residual Standard Error: 21.7 
Trace of smoother matrix: 5.53  (exact)

Control settings:
  span     :  0.5 
  degree   :  1 
  family   :  gaussian
  surface  :  interpolate     cell = 0.2
  normalize:  FALSE
 parametric:  FALSE
drop.square:  FALSE

现在我想提取这个模型的残差标准误差。我该如何提取它？我在模型对象的任何地方都找不到这个值 （即 21.7）。

> names(fit.loess[[i]])
 [1] "n"         "fitted"    "residuals" "enp"       "s"         "one.delta" "two.delta" "trace.hat"
 [9] "divisor"   "robust"    "pars"      "kd"        "call"      "terms"     "xnames"    "x"        
[17] "y"         "weights"

Answer 1

它是 loess 中 return 中的 s 元素。

> lo <- loess(mpg ~ wt, data=mtcars)
> print(lo)
#Call:
#loess(formula = mpg ~ wt, data = mtcars)
#
#Number of Observations: 32 
#Equivalent Number of Parameters: 5 
#Residual Standard Error: 2.711 
> lo$s
#[1] 2.711351

从 R 中的黄土回归结果中提取残差标准误差

Extract Residual Standard error from loess regression results in R

modeling

regression

r

loess