Description Usage Arguments Details Value Examples
This function creates a list of indices for a nonparametric bootstrap.
Corresponding to our ClustOmit statistic implemented in
clustomit
, we omit each cluster in turn and then sample from
the remaining clusters. We denote the number of groups as K
, which is
equal to nlevels(factor(y))
. Specifically, suppose that we omit the
kth group. That is, we ignore all of the observations corresponding to
group k. Then, we sample with replacement from each of the remaining
groups (i.e., every group except for group k), yielding a set of
bootstrap indices.
1 |
y |
a vector that denotes the grouping of each observation. It must be
coercible with |
num_reps |
the number of bootstrap replications to use for each group |
stratified |
Should the bootstrap replicates be stratified by cluster? By default, no. See Details. |
The bootstrap resampling employed randomly samples from the remaining
observations after a cluster is omitted. By default, we ensure that one
observation is selected from each remaining cluster to avoid potential
situations where the resampled data set contains multiple replicates of a
single observation. Optionally, by setting the stratified
argument to
TRUE
, we employ a stratified sampling scheme, where instead we sample
with replacement from each cluster. In this case, the number of observations
sampled from a cluster is equal to the number of observations originally
assigned to that cluster (i.e., its cluster size).
The returned list contains K * num_reps
elements.
Both resampling schemes ensure that we avoid errors when clustering, similar to this post on R Help: https://stat.ethz.ch/pipermail/r-help/2004-June/052357.html.
named list containing indices for each bootstrap replication
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 | set.seed(42)
# We use 4 clusters, each with up to 10 observations. The sample sizes are
# randomly chosen.
K <- 4
sample_sizes <- sample(10, K, replace = TRUE)
# Create the cluster labels, y.
y <- unlist(sapply(seq_len(K), function(k) {
rep(k, sample_sizes[k])
}))
# Use 20 reps per group.
boot_omit(y, num_reps = 20)
# Use the default number of reps per group.
boot_omit(y)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.