0

How can I extract the SE of delta (power function) and sigma of a gls objective? If I do the summary I have this output:

> summary(l3s)
Generalized least squares fit by REML
  Model: biom.kg ~ I(dbh^2 * h) 
  Data: ssf 
      AIC      BIC    logLik
  4495.75 4512.982 -2243.875

Variance function:
 Structure: Power of variance covariate
 Formula: ~I(dbh^2 * h) 
 Parameter estimates:
    power 
0.9383058 

Coefficients:
                  Value  Std.Error  t-value p-value
(Intercept)  -0.7649061 0.19920153 -3.83986   1e-04
I(dbh^2 * h)  0.0312859 0.00066244 47.22858   0e+00


 Correlation: 
             (Intr)
I(dbh^2 * h) -0.541


Standardized residuals:
        Min          Q1         Med          Q3         Max 
-1.98250304 -0.82199784 -0.07727941  0.69239313  3.05748659 

Residual standard error: 0.02026094 
Degrees of freedom: 551 total; 549 residual

Thank you.

Lstat
  • 1,450
  • 1
  • 12
  • 18
Dea
  • 1
  • 2

1 Answers1

0

Standard error of Power of variance covariate cannot be extracted as it is not computed by gls. Consider employing Monte Carlo simulation and extracting fit$modelStruct[2]. You can find sigma of your model in l3s$sigma.

Lstat
  • 1,450
  • 1
  • 12
  • 18