Inference for selector via bootstrapping — infer

Inference for selector via bootstrapping

Bootstrapping selection process

Usage

infer_boot(
  object,
  data,
  inference_target = c("selections", "all"),
  debias = TRUE,
  estimation_data = c("in-sample", "out-of-sample"),
  conf.level = 0.95,
  type = c("paired", "residual"),
  B = 250,
  n_cores = 4,
  ...
)

boot(
  object,
  data,
  B,
  inference_target = c("selections", "all"),
  debias = TRUE,
  estimation_data = c("in-sample", "out-of-sample"),
  conf.level,
  n_cores,
  ...
)

Arguments

object: a selector object
data: data must be passed to infer
inference_target: is inference requested on all or selected only
debias: should estimates be debiased (no, non-selections, or all)
estimation_data: within a bootstrap, should in-sample or out-of-sample residuals be used for estimation (defaults to FALSE)
conf.level: .95 by default
type: what type of bootstrap (currently only paired supported)
B: The number of bootstrap replicates.
n_cores: number of cores for parallel computation
...: any additional arguments to that can be passed to fitting engine

Value

inferrer s3 class with things like...

an inferrer object