Learn R Programming

rsample (version 0.1.0)

bootstraps: Bootstrap Sampling

Description

A bootstrap sample is a sample that is the same size as the original data set that is made using replacement. This results in analysis samples that have multiple replicates of some of the original rows of the data. The assessment set is defined as the rows of the original data that were not included in the bootstrap sample. This is often referred to as the "out-of-bag" (OOB) sample.

Usage

bootstraps(
  data,
  times = 25,
  strata = NULL,
  breaks = 4,
  pool = 0.1,
  apparent = FALSE,
  ...
)

Arguments

data

A data frame.

times

The number of bootstrap samples.

strata

A variable that is used to conduct stratified sampling. When not NULL, each bootstrap sample is created within the stratification variable. This could be a single character value or a variable name that corresponds to a variable that exists in the data frame.

breaks

A single number giving the number of bins desired to stratify a numeric stratification variable.

pool

A proportion of data used to determine if a particular group is too small and should be pooled into another group. We do not recommend decreasing this argument below its default of 0.1 because of the dangers of stratifying groups that are too small.

apparent

A logical. Should an extra resample be added where the analysis and holdout subset are the entire data set. This is required for some estimators used by the summary function that require the apparent error rate.

...

Not currently used.

Value

An tibble with classes bootstraps, rset, tbl_df, tbl, and data.frame. The results include a column for the data split objects and a column called id that has a character string with the resample identifier.

Details

The argument apparent enables the option of an additional "resample" where the analysis and assessment data sets are the same as the original data set. This can be required for some types of analysis of the bootstrap results. The strata argument is based on a similar argument in the random forest package were the bootstrap samples are conducted within the stratification variable. This can help ensure that the number of data points in the bootstrap sample is equivalent to the proportions in the original data set. (Strata below 10% of the total are pooled together by default.)

Examples

Run this code
# NOT RUN {
bootstraps(mtcars, times = 2)
bootstraps(mtcars, times = 2, apparent = TRUE)

library(purrr)
library(modeldata)
data(wa_churn)

set.seed(13)
resample1 <- bootstraps(wa_churn, times = 3)
map_dbl(resample1$splits,
        function(x) {
          dat <- as.data.frame(x)$churn
          mean(dat == "Yes")
        })

set.seed(13)
resample2 <- bootstraps(wa_churn, strata = churn, times = 3)
map_dbl(resample2$splits,
        function(x) {
          dat <- as.data.frame(x)$churn
          mean(dat == "Yes")
        })

set.seed(13)
resample3 <- bootstraps(wa_churn, strata = tenure, breaks = 6, times = 3)
map_dbl(resample3$splits,
        function(x) {
          dat <- as.data.frame(x)$churn
          mean(dat == "Yes")
        })
# }

Run the code above in your browser using DataLab