R: Auxiliary for controlling IB

ibControl {ib}

R Documentation

Auxiliary for controlling IB

Description

Auxiliary function for ib bias correction.

Usage

ibControl(
  tol = 1e-05,
  maxit = 25,
  verbose = FALSE,
  seed = 123L,
  H = 1L,
  constraint = TRUE,
  early_stop = FALSE,
  cens = FALSE,
  right = NULL,
  left = NULL,
  mis = FALSE,
  prop = NULL,
  out = FALSE,
  eps = NULL,
  G = NULL,
  func = function(x) rowMeans(x, na.rm = T),
  sim = NULL
)

Arguments

`tol`	positive convergence tolerance `\epsilon`. The `ib` procedure converges when `\|\|\hat{\theta}^{k+1}-\hat{\theta}^k\|\|_2/p<\epsilon`, where `p` is the dimension of `\theta`.
`maxit`	`integer` representing the maximal number of iterations.
`verbose`	if `TRUE`, it prints some output in the console at each iteration.
`seed`	`integer` to set the seed (see `Random`).
`H`	`integer` representing the number of bootstrap estimates (see `ib`).
`constraint`	if `TRUE` (default), constraint for `extra_param` is used in the iterative procedure (see 'Details' of `ib`).
`early_stop`	if `TRUE` (default is `FALSE`), the iterative procedure stops as soon as there is no improvment in the minimization of the objective function (see 'Details' of `ib`).
`cens`	if `TRUE` the simulated responses are censored according to `left` and `right` values.
`right`	`double` for right-censoring (only used if `cens=TRUE`).
`left`	`double` for left-censoring (only used if `cens=TRUE`).
`mis`	if `TRUE` the simulated responses have missing data at random.
`prop`	`double` between 0 and 1 representing the proportion of missing data (only used if `mis=TRUE`).
`out`	if `TRUE` the simulated responses are also generated with a contamination mechanism.
`eps`	`double` between 0 and 1 representing the proportion of outliers in the data (only used if `out=TRUE`).
`G`	a `function` to generate outliers. It takes only a sample size as argument.
`func`	a `function` to reduce the `H` bootstrap estimates (rowwise). By default, the average is computed. The user can supply a function. One could imagine using other function such as the median or a trimmed mean.
`sim`	a user-defined function for simulating responses (see 'Details')

Details

sim allows the user to provide its own function for generating responses. Currently it is only supported for generalized linear models with the prototype 'fun(object, control, extra_param, ...)' (see ib).

Value