This package provides jackknife resampling and estimation functions for Julia.
This package is registered in the General registry and can be installed from the Julia REPL with
julia> using Pkg
julia> Pkg.add("Jackknife")
or in the Pkg REPL mode with ]add Jackknife
.
None of the functions here are exported, so you'll have to call them with the prefix
Jackknife.
or explicitly import them.
Each function takes the following two arguments:
-
A point estimator, given as a
Function
. The function must return a scalar when passed a vector. -
A real-valued vector of length > 1.
leaveoneout(estimator, x)
Compute a vector of point estimates based on systematic subsamples of x
wherein
each index is omitted one at a time.
These are the "leave-one-out" estimates.
The resulting vector will have length length(x) - 1
.
variance(estimator, x)
The variance of the estimator computed using the jackknife technique.
bias(estimator, x)
The bias of the estimator computed using the jackknife technique.
estimate(estimator, x)
The bias-corrected jackknife estimate of the parameter.