bootstrap_pffr: Perform bootstrap for smooth effects in 'pffr' models

In bauer-alex/FoSIntro: Convenience Functions for Semiparametric Function-on-Scalar Regression

Description

With this function you can perform a parametric or nonparametric bootstrap for a function-on-scalar model fitted with pffr. Based on the results, confidence intervals (CIs) for smooth effects can be calculated using link{calc_bootstrapCIs}.

Usage

bootstrap_pffr(type = "parametric", B = 1000, formula, data,
  model = NULL, param_yvar = NULL, cores = 1,
  param_simFun = "simulate", param_yMinValue = NULL, log_file = NULL,
  ...)

Arguments

type	One of "parametric" or "nonparametric", specifying the type of bootstrap to be used. If type = "parametric" you also have to specify the model argument.
B	Number of bootstrap samples, defaults to 1000
formula	See formula argument of pffr
data	See data argument of pffr
model	Function-on-scalar model fitted with pffr used for parametric bootstrapping. Only used if type=='parametric'.
param_yvar	Name of response variable used in model and present in data. Only used if type=='parametric'.
cores	Number of cores to use for parallel processing. Possible for both Linux-based systems and Windows.
param_simFun	Specifies the simulation function to be used for parametric bootstrapping. See simulate_pffr.
param_yMinValue	Minimum value to which the bootstrapped response values are set using a parametric bootstrap. See simulate_pffr.
log_file	Optional filename of a log file where progress of parallel processing should printed to. Defaults to console output (only on Linux-based systems). Only used if cores>1.
...	Further arguments passed to pffr

bauer-alex/FoSIntro documentation built on Feb. 11, 2022, 8:33 a.m.