pbcm.di: Data-informed Parametric Bootstrap Cross-fitting
In hkauhanen/pbcm: Parametric Bootstrap Cross-Fitting Method

Description Usage Arguments Details Value Author(s) References Examples

View source: R/pbcm.di.R

The data-informed Parametric Bootstrap Cross-fitting Method (PBCM) generates synthetic data from two models of a phenomenon parameterized by fits to an empirical dataset, and then cross-fits the models to these data. The result is two distributions of the goodness of fit metric Δ GoF = GoF_1 - GoF_2, where GoF_1 is the fit of model 1 and GoF_2 the fit of model 2.

pbcm.di(
  data,
  fun1,
  fun2,
  genfun1,
  genfun2,
  reps,
  args1 = NULL,
  args2 = NULL,
  genargs1 = NULL,
  genargs2 = NULL,
  print_genargs = TRUE,
  nonparametric_bootstrap = TRUE,
  verbose = TRUE,
  GoFname = "GoF"
)

`data`	Data frame
`fun1`	First modelling function
`fun2`	Second modelling function
`genfun1`	Generator function for first model
`genfun2`	Generator function for second model
`reps`	Number of Monte Carlo repetitions
`args1`	List of arguments passed to `fun1`
`args2`	List of arguments passed to `fun2`
`genargs1`	List of arguments passed to `genfun1`
`genargs2`	List of arguments passed to `genfun2`
`print_genargs`	Whether the generator argument values should be included in output (see Details)
`nonparametric_bootstrap`	Whether `data` should be nonparametrically bootstrapped before the parametric bootstrap
`verbose`	If `TRUE`, a progress bar is printed to the console and warnings are issued
`GoFname`	Name of the element returned by `fun1` and `fun2` holding the goodness of fit; see Details

Functions fun1 and fun2 must take data as an argument in addition to any arguments specified in args1 and args2. Moreover, these functions must return a list with at least one element carrying the goodness of fit; the name of this element may be specified through the GoFname argument, by default the string "GoF" is assumed. Functions genfun1 and genfun2 must take an argument named model (the output of fun1 and fun2).

A data frame in long format with the following columns:

rep: Monte Carlo repetition number
generator: Generating model
GoF1: Goodness of fit of model 1
GoF2: Goodness of fit of model 2
DeltaGoF: Equals GoF1 - GoF2

In addition to these columns, if print_genargs == TRUE, each argument in the lists genargs1 and genargs2 is included as a column of its own, with the argument's name prefixed by "genargs1_" or "genargs2_".

Henri Kauhanen

Wagenmakers, E.-J., Ratcliff, R., Gomez, P. & Iverson, G. J. (2004) Assessing model mimicry using the parametric bootstrap. Journal of Mathematical Psychology, 48(1), 28–50. https://doi.org/10.1016/j.jmp.2003.11.004

x <- seq(from=0, to=1, length.out=100)
mockdata <- data.frame(x=x, y=x + rnorm(100, 0, 0.5))

myfitfun <- function(data, p) {
  res <- nls(y~a*x^p, data, start=list(a=1.1))
  list(a=coef(res), GoF=deviance(res))
}

mygenfun <- function(model, p) { 
  x <- seq(from=0, to=1, length.out=100)
  y <- model$a*x^p + rnorm(100, 0, 0.5)
  data.frame(x=x, y=y)
}

pbcm.di(data=mockdata, fun1=myfitfun, fun2=myfitfun, genfun1=mygenfun,
        genfun2=mygenfun, reps=20, args1=list(p=1), args2=list(p=2), 
        genargs1=list(p=1), genargs2=list(p=2))