sampcla: Within-class sampling

In rchemo: Dimension Reduction, Regression and Discrimination for Chemometrics

Within-class sampling

Description

The function divides a datset in two sets, "train" vs "test", using a stratified sampling on defined classes.

If argument y = NULL (default), the sampling is random within each class. If not, the sampling is systematic (regular grid) within each class over the quantitative variable y.

Usage

sampcla(x, y = NULL, m)

Arguments

x	A vector (length m) defining the class membership of the observations.
y	A vector (length m) defining the quantitative variable for the systematic sampling. If NULL (default), the sampling is random within each class.
m	Either an integer defining the equal number of test observation(s) to select per class, or a vector of integers defining the numbers to select for each class. In the last case, vector m must have a length equal to the number of classes present in x, and be ordered in the same way as the ordered class membership.

Value

train	Indexes (i.e. position in x) of the selected observations, for the training set.
test	Indexes (i.e. position in x) of the selected observations, for the test set.
lev	classes
ni	number of observations in each class

Note

The second example is a representative stratified sampling from an unsupervised clustering.

References

Naes, T., 1987. The design of calibration in near infra-red reflectance analysis by clustering. Journal of Chemometrics 1, 121-134.

Examples

## EXAMPLE 1

x <- sample(c(1, 3, 4), size = 20, replace = TRUE)
table(x)

sampcla(x, m = 2)
s <- sampcla(x, m = 2)$test
x[s]

sampcla(x, m = c(1, 2, 1))
s <- sampcla(x, m = c(1, 2, 1))$test
x[s]

y <- rnorm(length(x))
sampcla(x, y, m = 2)
s <- sampcla(x, y, m = 2)$test
x[s]

## EXAMPLE 2

data(cassav)
X <- cassav$Xtrain
y <- cassav$ytrain
N <- nrow(X)

fm <- pcaeigenk(X, nlv = 10)
z <- stats::kmeans(x = fm$T, centers = 3, nstart = 25, iter.max = 50)
x <- z$cluster
z <- table(x)
z
p <- c(z) / N
p

psamp <- .20
m <- round(psamp * N * p)
m

random_sampling <- sampcla(x, m = m)
s <- random_sampling$test
table(x[s])

Systematic_sampling_for_y <- sampcla(x, y, m = m)
s <- Systematic_sampling_for_y$test
table(x[s])

rchemo documentation built on Sept. 11, 2024, 8:05 p.m.