popApplyTwDEMC: popApplyTwDEMC

In twDEMC: parallel DEMC

Description

Applying a function across all chains of one population for each case.

Usage

popApplyTwDEMC(x, nPop, FUN, ...)

Arguments

x	a matrix with columns chains or array with last dimension chain
nPop	number of populations
FUN	function to apply to population submatrix
...	further arguemtns to FUN

Value

array with last dimenstion correponding to population

Author(s)

Thomas Wutzler

See Also

popMeansTwDEMC subChains.twDEMC

Examples

data(twdemcEx1)
ex1c <- concatPops(twdemcEx1)
# mean logDen for each case, i.e. step, by population
nPop=getNPops(ex1c)
#mtrace(popApplyTwDEMC)
# applied to a matrix: pops in columns
popApplyTwDEMC( ex1c$logDen[,1,], nPop=nPop, apply, 1, mean )
# applied to a 3d array: pops in 3rd dimension
tmp <- popApplyTwDEMC( ex1c$logDen, nPop=nPop, apply, 1:2, mean )
str(tmp)

# stack logDen of block 1 for each population
tmp <- popApplyTwDEMC( ex1c$logDen[,1,], nPop=nPop, as.vector )
str(tmp)
#stack param columns by population
str(ex1c$parms)	#26 cases, 2 parameters, 2*4=8 chains
tmp <- popApplyTwDEMC( ex1c$parms, nPop=nPop, function(x){ 
                abind::abind(twMisc::twListArrDim(x),along=1) })
all.equal( c(nrow(ex1c$parms)*4, 2,2), dim(tmp))

twDEMC documentation built on May 2, 2019, 5:38 p.m.