nlme.diag: Diagnostic plots for nlme

In qPharmetra/qpToolkit: qPharmetra R Tools

Diagnostic plots for nlme

Description

Given an R nlme model object, create two diagnostic plots of the eta(s) in the model

Usage

## S3 method for class 'diag'
nlme(
  obj,
  subset.modeldata,
  xvar = "time",
  xvar.label = NULL,
  nx = 8,
  output = FALSE,
  print.eta.norm = TRUE,
  print.eta.v.var = TRUE,
  asp.eta.norm = 1,
  asp.eta.v.var = 1
)

Arguments

obj	nlme model object
subset.modeldata	character string to subset the data
xvar	the column name of the independent variable
xvar.label	label for independent variable, defaulting to NULL (no label)
nx	number of bins of xvar to bin across
output	logical determining if the plotting data should be outputted
print.eta.norm	logical determining if a standard normal Q-Q plot of the eta estimates will be plotted print.eta.v.var
print.eta.v.var	logical determining if a distribution of etas by binned values of xvar will be plotted
asp.eta.norm	aspect value for the QQ plot
asp.eta.v.var	aspect vlaue of the plot for eta vs variables

Value

Diagnostic plots

See Also

nlme.run, nlme.predict

Examples

## define modeling function
## adapted from pk.1comp.1abs to make it modeling-ready
library(nlme)
pkpdData = example.pkpdData()
PK.1comp.1abs =
function(dose, tob, cl, v, ka){
  kel = cl / v
  dose * ka/v/(ka-kel) * (exp(-kel*tob) - exp(-ka*tob))
}

## fit 1 comp PK with 1st order absorption
fit.nlme.1 = nlme.run(value ~ PK.1comp.1abs(dose, time, cl*exp(cl.eta), v*exp(v.eta), ka),
                     data = subset(pkpdData, type == "PK" & dose> 0 & value > 0.1),
                     groups = ~ id,
                     fixed = cl + v + ka ~ 1,
                     random = pdDiag(list(cl.eta~1,v.eta ~ 1)),
                     start = c(cl = 1, v = 5, ka = 1),
                     reference = 3,
                     problem = "1comp.1abs eta(CL)"
)

summary(fit.nlme.1$object)
nlme.diag(fit.nlme.1$object)
# note here we refer to the $object, given the model was created with nlme.run()

qPharmetra/qpToolkit documentation built on May 24, 2023, 8:52 a.m.