balseq: Sequential balanced sampling

In RJauslin/StratifiedSampling: Different Methods for Stratified Sampling

Sequential balanced sampling

Description

Selects at the same time a well-spread and a balanced sample using a sequential implementation.

Usage

balseq(pik, Xaux, Xspread = NULL, rord = TRUE)

Arguments

pik	A vector of inclusion probabilities.
Xaux	A matrix of auxiliary variables. The matrix must contains the pik vector to have fixed sample size.
Xspread	An optional matrix of spatial coordinates.
rord	A logical variable that specify if reordering is applied. Default TRUE.

Details

The function selects a sample using a sequential algorithm. At the same time, it respects the balancing equations (Xaux) and select a well-spread sample (Xspread). Algorithm uses a linear program to satisfy the constraints.

Value

Return the selected indices in 1,2,...,N

See Also

BalancedSampling:lcube, sampling:samplecube.

Examples

N <- 100
n <- 10
p <- 10

pik <- rep(n/N,N)

Xaux <- array(rnorm(N*p,3,1),c(N,p))

Xspread <- cbind(runif(N),runif(N)) 
Xaux <- cbind(pik,Xaux)

s <- balseq(pik,Xaux)
colSums(Xaux[s,]/as.vector(pik[s]))
colSums(Xaux)


s <- balseq(pik,Xaux,Xspread)
colSums(Xaux[s,]/as.vector(pik[s]))
colSums(Xaux)

RJauslin/StratifiedSampling documentation built on Feb. 5, 2025, 4:18 a.m.