dynmodel: Fit a non-population dynamic model

In nlmixr: Nonlinear Mixed Effects Models in Population PK/PD

Fit a non-population dynamic model

Description

Fit a non-population dynamic model

Usage

dynmodel(
  system,
  model,
  inits,
  data,
  fixPars = NULL,
  nlmixrObject = NULL,
  control = list(),
  ...
)

Arguments

system	RxODE object. See RxODE for more details.
model	Error model.
inits	Initial values of system parameters.
data	Dataset to estimate. Needs to be RxODE compatible in EVIDs.
fixPars	Fixed system parameters. Default is NULL.
nlmixrObject	nlmixr object. See nlmixr for more details. Default is NULL.
control	Control options for dynmodel dynmodelControl .
...	Other parameters (ignored)

Value

A dynmodel fit object

Author(s)

Wenping Wang, Mason McComb and Matt Fidler

Examples

# dynmodel example --------------------------------------------------------
ode <- "
      kel = CL/V;
      d/dt(X) = -kel*X;
      C=X/V;
      PRED = C
      "
ode_system <- RxODE(model = ode)
model_error_structure <- cp ~ C + add(0.01) + prop(0.01)
inits <- c(CL = 1, V = 10)
control <- dynmodelControl(method = "Nelder-Mead")
fit <-
  try(dynmodel(
    system = ode_system,
    model = model_error_structure,
    data = Bolus_1CPT,
    inits = inits,
    control = control
  ))

# nlmixr model example ----------------------------------------------------------
model_onecmt_bolus <- function() {
  ini({
    CL <- c(0, 5, 10) # Clearance (L/hr)
    V <- c(0, 50, 100) # Volume of Distribution
    prop.err <- c(0, 0.01, 1)
  })
  model({
    kel <- CL / V
    d / dt(X) <- -kel * X
    cp <- X / V
    cp ~ prop(prop.err)
  })
}

# note on some platforms this fit is not successful
fit <- try(nlmixr(object = model_onecmt_bolus, data = Bolus_1CPT, est = "dynmodel"))

if (inherits(fit, "nlmixrDynmodel")) {
 as.dynmodel(fit)
}

# method = "focei" is slightly more flexible and well tested

fit <- try(nlmixr(object = model_onecmt_bolus, data = Bolus_1CPT, est = "focei"))

nlmixr documentation built on March 27, 2022, 5:05 p.m.