From cf54ccca37d27480dbf8d59eb027300518f7ad75 Mon Sep 17 00:00:00 2001
From: Johannes Ranke Last change 20 April 2023
-(rebuilt 2023-04-20)
+(rebuilt 2023-05-19)
Source: vignettes/web_only/multistart.rmd
multistart.rmd
-f_saem_reduced <- update(f_saem_full, no_random_effect = "log_k2")
+f_saem_reduced <- stats::update(f_saem_full, no_random_effect = "log_k2")
illparms(f_saem_reduced)
f_saem_reduced_multi <- multistart(f_saem_reduced, n = 16, cores = 16)
parplot(f_saem_reduced_multi, lpos = "topright", ylim = c(0.5, 2))
We can use the anova
method to compare the models.
## Data: 155 observations of 1 variable(s) grouped in 6 datasets
##
-## npar AIC BIC Lik
-## f_saem_reduced 9 663.73 661.86 -322.86
-## best(f_saem_reduced_multi) 9 663.69 661.82 -322.85
-## f_saem_full 10 669.77 667.69 -324.89
-## best(f_saem_full_multi) 10 665.56 663.48 -322.78
-The reduced model gives the lowest information criteria and similar -likelihoods as the best variant of the full model. The multistart method -leads to a much lower improvement of the likelihood for the reduced -model, indicating that it converges faster.
+## npar AIC BIC Lik Chisq Df Pr(>Chisq) +## f_saem_reduced 9 663.67 661.80 -322.84 +## best(f_saem_reduced_multi) 9 663.65 661.78 -322.82 0.0219 0 +## f_saem_full 10 670.09 668.01 -325.05 0.0000 1 1 +## best(f_saem_full_multi) 10 665.61 663.52 -322.80 4.4870 0 +The reduced model results in lower AIC and BIC values, so it is +clearly preferable. Using multiple starting values gives a large +improvement in case of the full model, because it is less well-defined, +which impedes convergence. For the reduced model, using multiple +starting values only results in a small improvement of the model +fit.