bootstrap: Generate bootstrap replicates

In jrnold/resamplr: Resampling methods

Description

Generate bootstrap replicates. In addition to the ordinary bootstrap, it supports weighted bootstrap (weights), Bayesian bootstrap (bayes), cluster bootstrap (if data is a grouped data frame and groups = TRUE), and stratified bootstrap (if data is grouped data frame and stratify = TRUE).

Usage

bootstrap(data, ...)

## S3 method for class 'data.frame'
bootstrap(data, R = 1L, weights = NULL,
  bayes = FALSE, m = NULL, ...)

## S3 method for class 'grouped_df'
bootstrap(data, R = 1L, stratify = FALSE,
  groups = TRUE, weights = NULL, bayes = FALSE, weight_groups = TRUE,
  weight_within = TRUE, ...)

Arguments

data	A data frame
...	Passed to methods
R	Number of replicates to sample
weights	Observation level weights. This must be the name of a column in data.
bayes	If TRUE, a Bayesian bootstrap is used.
m	Number of observations in the replicates.
stratify	Resample within groups (stratified bootstrap)
groups	Resample groups (clustered bootstrap)
weight_groups	If TRUE and groups = TRUE, then sample groups using group-level weights. If weights is NULL, then weight groups by the number of observations in each. If weights, is not-NULL, then use the sum of the observation level weights in each group as the group-level weight.
weight_within	If TRUE and stratify = TRUE, then use observation level weights to sample observations within each group.

Value

A data frame with R rows and the following columns:

train

A list of resample objects. Training sets.

test

A list of resample objects. Test sets.

.id

An integer vector of identifiers

Methods (by class)

• data.frame: Bootstrap rows in a data frame.

• grouped_df: Bootstraps a grouped data data frame. Allows for bootstrapping groups (clustered bootstrap) if groups = TRUE, and bootstrapping within groups (stratified bootstrap) if stratify = TRUE.

References

• Angelo Canty and Brian Ripley (2016). boot: Bootstrap R (S-Plus) Functions. R package version 1.3-18.

• Davison, A. C. & Hinkley, D. V. (1997) Bootstrap Methods and Their Applications. Cambridge University Press, Cambridge. ISBN 0-521-57391-2

See Also

The boot function boot which is the canonical R bootstrap implementation. This function produces slightly different output and has more options than the modelr function bootstrap.

Examples

# example from modelr::bootstrap
library("purrr")
library("dplyr")
boot <- bootstrap(mtcars, 100)

models <- map(boot$sample, ~ lm(mpg ~ wt, data = .))
tidied <- map_df(models, broom::tidy, .id = "id")

hist(subset(tidied, term == "wt")$estimate)
hist(subset(tidied, term == "(Intercept)")$estimate)

jrnold/resamplr documentation built on May 20, 2019, 1:05 a.m.