diff options
Diffstat (limited to 'vignettes/web_only/dimethenamid_2018.rmd')
| -rw-r--r-- | vignettes/web_only/dimethenamid_2018.rmd | 57 |
1 files changed, 40 insertions, 17 deletions
diff --git a/vignettes/web_only/dimethenamid_2018.rmd b/vignettes/web_only/dimethenamid_2018.rmd index d3541a34..30325044 100644 --- a/vignettes/web_only/dimethenamid_2018.rmd +++ b/vignettes/web_only/dimethenamid_2018.rmd @@ -1,7 +1,7 @@ --- title: Example evaluations of the dimethenamid data from 2018 author: Johannes Ranke -date: Last change 23 June 2021, built on `r format(Sys.Date(), format = "%d %b %Y")` +date: Last change 27 July 2021, built on `r format(Sys.Date(), format = "%d %b %Y")` output: html_document: toc: true @@ -163,8 +163,10 @@ tendency of the algorithm to try parameter combinations unsuitable for fitting these data. ```{r f_parent_nlme, warning = FALSE} +library(nlme) f_parent_nlme_sfo_const <- nlme(f_parent_mkin_const["SFO", ]) -#f_parent_nlme_dfop_const <- nlme(f_parent_mkin_const["DFOP", ]) # error +#f_parent_nlme_dfop_const <- nlme(f_parent_mkin_const["DFOP", ]) +# maxIter = 50 reached f_parent_nlme_sfo_tc <- nlme(f_parent_mkin_tc["SFO", ]) f_parent_nlme_dfop_tc <- nlme(f_parent_mkin_tc["DFOP", ]) ``` @@ -180,10 +182,10 @@ used for these attempts can be made visible below. f_parent_nlme_sfo_const_logchol <- nlme(f_parent_mkin_const["SFO", ], random = pdLogChol(list(DMTA_0 ~ 1, log_k_DMTA ~ 1))) anova(f_parent_nlme_sfo_const, f_parent_nlme_sfo_const_logchol) # not better -f_parent_nlme_dfop_tc_logchol <- update(f_parent_nlme_dfop_tc, - random = pdLogChol(list(DMTA_0 ~ 1, log_k1 ~ 1, log_k2 ~ 1, g_qlogis ~ 1))) +#f_parent_nlme_dfop_tc_logchol <- update(f_parent_nlme_dfop_tc, +# random = pdLogChol(list(DMTA_0 ~ 1, log_k1 ~ 1, log_k2 ~ 1, g_qlogis ~ 1))) # using log Cholesky parameterisation for random effects (nlme default) does -# not converge and gives lots of warnings about the LME step not converging +# not converge here and gives lots of warnings about the LME step not converging ``` The model comparison function of the nlme package can directly be applied @@ -221,7 +223,7 @@ The convergence plot for the SFO model using constant variance is shown below. ```{r f_parent_saemix_sfo_const, results = 'hide'} library(saemix) -f_parent_saemix_sfo_const <- saem(f_parent_mkin_const["SFO", ], quiet = TRUE, +f_parent_saemix_sfo_const <- mkin::saem(f_parent_mkin_const["SFO", ], quiet = TRUE, transformations = "saemix") plot(f_parent_saemix_sfo_const$so, plot.type = "convergence") ``` @@ -230,18 +232,19 @@ Obviously the default number of iterations is sufficient to reach convergence. This can also be said for the SFO fit using the two-component error model. ```{r f_parent_saemix_sfo_tc, results = 'hide'} -f_parent_saemix_sfo_tc <- saem(f_parent_mkin_tc["SFO", ], quiet = TRUE, +f_parent_saemix_sfo_tc <- mkin::saem(f_parent_mkin_tc["SFO", ], quiet = TRUE, transformations = "saemix") plot(f_parent_saemix_sfo_tc$so, plot.type = "convergence") ``` When fitting the DFOP model with constant variance, parameter convergence -is not as unambiguous. Therefore, the number of iterations in the first +is not as unambiguous (see the failure of nlme with the default number of +iterations above). Therefore, the number of iterations in the first phase of the algorithm was increased, leading to visually satisfying convergence. ```{r f_parent_saemix_dfop_const, results = 'hide'} -f_parent_saemix_dfop_const <- saem(f_parent_mkin_const["DFOP", ], quiet = TRUE, +f_parent_saemix_dfop_const <- mkin::saem(f_parent_mkin_const["DFOP", ], quiet = TRUE, control = saemixControl(nbiter.saemix = c(800, 200), print = FALSE, save = FALSE, save.graphs = FALSE, displayProgress = FALSE), transformations = "saemix") @@ -250,11 +253,11 @@ plot(f_parent_saemix_dfop_const$so, plot.type = "convergence") The same applies to the case where the DFOP model is fitted with the two-component error model. Convergence of the variance of k2 is enhanced -by using the two-component error, it remains pretty stable already after 200 +by using the two-component error, it remains more or less stable already after 200 iterations of the first phase. ```{r f_parent_saemix_dfop_tc_moreiter, results = 'hide'} -f_parent_saemix_dfop_tc_moreiter <- saem(f_parent_mkin_tc["DFOP", ], quiet = TRUE, +f_parent_saemix_dfop_tc_moreiter <- mkin::saem(f_parent_mkin_tc["DFOP", ], quiet = TRUE, control = saemixControl(nbiter.saemix = c(800, 200), print = FALSE, save = FALSE, save.graphs = FALSE, displayProgress = FALSE), transformations = "saemix") @@ -306,6 +309,7 @@ First, the focei algorithm is used for the four model combinations and the goodness of fit of the results is compared. ```{r f_parent_nlmixr_focei, results = "hide", message = FALSE, warning = FALSE} +library(nlmixr) f_parent_nlmixr_focei_sfo_const <- nlmixr(f_parent_mkin_const["SFO", ], est = "focei") f_parent_nlmixr_focei_sfo_tc <- nlmixr(f_parent_mkin_tc["SFO", ], est = "focei") f_parent_nlmixr_focei_dfop_const <- nlmixr(f_parent_mkin_const["DFOP", ], est = "focei") @@ -317,7 +321,14 @@ AIC(f_parent_nlmixr_focei_sfo_const$nm, f_parent_nlmixr_focei_sfo_tc$nm, f_parent_nlmixr_focei_dfop_const$nm, f_parent_nlmixr_focei_dfop_tc$nm) ``` -The AIC values are very close to the ones obtained with nlme. +The AIC values are very close to the ones obtained with nlme which are repeated below +for convenience. + +```{r AIC_parent_nlme_rep} +AIC( + f_parent_nlme_sfo_const, f_parent_nlme_sfo_tc, f_parent_nlme_dfop_tc +) +``` Secondly, we use the SAEM estimation routine and check the convergence plots for SFO with constant variance @@ -333,7 +344,7 @@ for SFO with two-component error ```{r f_parent_nlmixr_saem_sfo_tc, results = "hide", warning = FALSE, message = FALSE} f_parent_nlmixr_saem_sfo_tc <- nlmixr(f_parent_mkin_tc["SFO", ], est = "saem", control = nlmixr::saemControl(logLik = TRUE)) -nlmixr::traceplot(f_parent_nlmixr_saem_sfo_tc$nm) +traceplot(f_parent_nlmixr_saem_sfo_tc$nm) ``` For DFOP with constant variance, the convergence plots show considerable instability @@ -343,7 +354,7 @@ the number of parallel chains for the first phase of algorithm. ```{r f_parent_nlmixr_saem_dfop_const, results = "hide", warning = FALSE, message = FALSE} f_parent_nlmixr_saem_dfop_const <- nlmixr(f_parent_mkin_const["DFOP", ], est = "saem", control = nlmixr::saemControl(logLik = TRUE, nBurn = 1000), nmc = 15) -nlmixr::traceplot(f_parent_nlmixr_saem_dfop_const$nm) +traceplot(f_parent_nlmixr_saem_dfop_const$nm) ``` For DFOP with two-component error, the same increase in iterations and parallel @@ -354,7 +365,7 @@ erratic convergence, so this may not be necessary to this degree. ```{r f_parent_nlmixr_saem_dfop_tc, results = "hide", warning = FALSE, message = FALSE} f_parent_nlmixr_saem_dfop_tc <- nlmixr(f_parent_mkin_tc["DFOP", ], est = "saem", control = nlmixr::saemControl(logLik = TRUE, nBurn = 1000, nmc = 15)) -nlmixr::traceplot(f_parent_nlmixr_saem_dfop_tc$nm) +traceplot(f_parent_nlmixr_saem_dfop_tc$nm) ``` The AIC values are internally calculated using Gaussian quadrature. For an @@ -366,8 +377,20 @@ AIC(f_parent_nlmixr_saem_sfo_const$nm, f_parent_nlmixr_saem_sfo_tc$nm, f_parent_nlmixr_saem_dfop_const$nm, f_parent_nlmixr_saem_dfop_tc$nm) ``` - - +The following table gives the AIC values obtained with the three packages. + +```{r AIC_all} +AIC_all <- data.frame( + nlme = c(AIC(f_parent_nlme_sfo_const), AIC(f_parent_nlme_sfo_tc), NA, AIC(f_parent_nlme_dfop_tc)), + nlmixr_focei = sapply(list(f_parent_nlmixr_focei_sfo_const$nm, f_parent_nlmixr_focei_sfo_tc$nm, + f_parent_nlmixr_focei_dfop_const$nm, f_parent_nlmixr_focei_dfop_tc$nm), AIC), + saemix = sapply(list(f_parent_saemix_sfo_const$so, f_parent_saemix_sfo_tc$so, + f_parent_saemix_dfop_const$so, f_parent_saemix_dfop_tc_moreiter$so), AIC), + nlmixr_saem = sapply(list(f_parent_nlmixr_saem_sfo_const$nm, f_parent_nlmixr_saem_sfo_tc$nm, + f_parent_nlmixr_saem_dfop_const$nm, f_parent_nlmixr_saem_dfop_tc$nm), AIC) +) +kable(AIC_all) +``` # References |