Generates wild bootstrap replicates of a vector or matrix of residuals
by multiplying each observation by a random Rademacher weight (+1 or -1).
Usage
wild_bootstrap(x, num_boots = 100, parallel = FALSE, num_cores = 2)
Arguments
- x
Numeric vector or matrix of residuals.
- num_boots
Integer number of bootstrap replicates.
- parallel
Parallelize computation? TRUE or FALSE.
- num_cores
Number of cores.
Value
A list of numeric matrices, each one a wild bootstrap replicate.
Details
The wild bootstrap is often used to resample regression or model residuals
when heteroskedasticity or other non-i.i.d. errors are present.
Each replicate is constructed by multiplying every observation by +1 or -1,
where the signs are drawn randomly with equal probability.
References
A. Colin Cameron & Jonah B. Gelbach & Douglas L. Miller, 2008.
"Bootstrap-Based Improvements for Inference with Clustered Errors",
The Review of Economics and Statistics, MIT Press, vol. 90(3), pages
414-427, August.
Examples
set.seed(123)
resids <- rnorm(100)
boot_reps <- wild_bootstrap(resids, num_boots = 5)
length(boot_reps) # 5 replicates
#> [1] 5
dim(boot_reps[[1]]) # 100 x 1 matrix
#> [1] 100 1
