smcUtils (version 0.2.2)

resample: Resample

Description

A set of resampling functions with unbiased number of replicates.

Usage

resample(weights, num.samples=length(weights), method = c("stratified","residual","multinomial","systematic","branching"), nonuniformity = c("none","ess","cov","entropy"), threshold = 0.5, rrf = "stratified", engine="R", log=TRUE, normalized=FALSE)

Arguments

weights

a vector of weights, possibly logged, unnormalized, or both

num.samples

a scalar indicating the number of samples to return (for `branching.resample', `num.samples' is the expected number of samples as the actual number is random)

method

a character string indicating the resampling method to use. One of `"multinomial"' (default), `"residual"', `"stratified"', `"systematic"', or `"branching"', can be abbreviated.

nonuniformity

a character string indicating which hueristic to use for measuring weight nonuniformity. One of `"none"' (default), `"ess"', `"cov"', or `"entropy"', can be abbreviated. If "none", resampling will always be performed.

threshold

a scalar in [0,1] indicating when to resample. If `nonuniformity="none"', resampling is always performed regardless of threshold. If `nonuniformity="ess"' or "entropy", resampling is performed when `nonuniformity/num.samples threshold'.

rrf

for residual resampling, the resampling function to use on the residual

engine

run using "R" or "C" code

log

if TRUE, the weights are assumed to be logged

normalized

if TRUE, the weights are assumed to be normalized and no error checking is done

Value

weights

the component weights

indices

an integer vector containing the indices of resampled components. If no resampling was performed, then `indices=1:length(weights)'.

See Also

resampling

Examples

ws = dnorm(0:5, log=TRUE)
resample(ws)
resample(ws,13)
resample(ws,method="residual")
resample(ws,method="residual",rrf="stratified")
resample(ws,17,"stratified","ess",0.5)