kenStone: Kennard-Stone algorithm for calibration sampling

In prospectr: Miscellaneous Functions for Processing and Sample Selection of Spectroscopic Data

Kennard-Stone algorithm for calibration sampling

Description

\loadmathjax

Select calibration samples from a large multivariate data using the Kennard-Stone algorithm

Usage

kenStone(X, k, metric = "mahal", pc, group,
         .center = TRUE, .scale = FALSE, init = NULL)

Arguments

X	a numeric matrix.
k	number of calibration samples to be selected.
metric	distance metric to be used: 'euclid' (Euclidean distance) or 'mahal' (Mahalanobis distance, default).
pc	optional. If not specified, distance are computed in the Euclidean space. Alternatively, distance are computed in the principal component score space and pc is the number of principal components retained. If pc < 1, the number of principal components kept corresponds to the number of components explaining at least (pc * 100) percent of the total variance.
group	An optional factor (or vector that can be coerced to a factor by as.factor) of length equal to nrow(X), giving the identifier of related observations (e.g. samples of the same batch of measurements, samples of the same origin, or of the same soil profile). When one observation is selected by the procedure all observations of the same group are removed together and assigned to the calibration set. This allows to select calibration points that are independent from the remaining points.
.center	logical value indicating whether the input matrix should be centered before Principal Component Analysis. Default set to TRUE.
.scale	logical value indicating whether the input matrix should be scaled before Principal Component Analysis. Default set to FALSE.
init	(optional) a vector of integers indicating the indices of the observations/rows that are to be used as observations that must be included at the first iteration of the search process. Default is NULL, i.e. no fixed initialization. The function will take by default the two most distant observations. If the group argument is used, then all the observations in the groups covered by the init observations will be also included in the init subset.

Details

The Kennard–Stone algorithm allows to select samples with a uniform distribution over the predictor space (Kennard and Stone, 1969). It starts by selecting the pair of points that are the farthest apart. They are assigned to the calibration set and removed from the list of points. Then, the procedure assigns remaining points to the calibration set by computing the distance between each unassigned points \mjeqni_0i_0 and selected points \mjeqnii and finding the point for which:

\mjdeqn

d_selected = \max\limits_i_0(\min\limits_i(d_i,i_0))d_sel ected = \max_i_0(\min_i(d_i,i0))

This essentially selects point \mjeqni_0i_0 which is the farthest apart from its closest neighbors \mjeqnii in the calibration set. The algorithm uses the Euclidean distance to select the points. However, the Mahalanobis distance can also be used. This can be achieved by performing a PCA on the input data and computing the Euclidean distance on the truncated score matrix according to the following definition of the Mahalanobis \mjeqnHH distance:

\mjdeqn

H_ij^2 = \sum_a=1^A (\hat t_ia - \hat t_ja)^2 / \hat \lambda_aH_ij^2 = sum_a=1^A (hat t_ia - hat t_ja)^2 / hat lambda_a

where \mjeqn\hat t_iahatt_ia is the \mjeqna^tha^th principal component score of point \mjeqnii, \mjeqn\hat t_jahatt_ja is the corresponding value for point \mjeqnjj, \mjeqn\hat \lambda_ahat lambda_a is the eigenvalue of principal component \mjeqnaa and \mjeqnAA is the number of principal components included in the computation.

Value

a list with the following components:

• model: numeric vector giving the row indices of the input data selected for calibration

• test: numeric vector giving the row indices of the remaining observations

• pc: if the pc argument is specified, a numeric matrix of the scaled pc scores

Author(s)

Antoine Stevens & Leonardo Ramirez-Lopez with contributions from Thorsten Behrens and Philipp Baumann

References

Kennard, R.W., and Stone, L.A., 1969. Computer aided design of experiments. Technometrics 11, 137-148.

See Also

duplex, shenkWest, naes, honigs

Examples

data(NIRsoil)
sel <- kenStone(NIRsoil$spc, k = 30, pc = .99)
plot(sel$pc[, 1:2], xlab = "PC1", ylab = "PC2")
# points selected for calibration
points(sel$pc[sel$model, 1:2], pch = 19, col = 2)
# Test on artificial data
X <- expand.grid(1:20, 1:20) + rnorm(1e5, 0, .1)
plot(X, xlab = "VAR1", ylab = "VAR2")
sel <- kenStone(X, k = 25, metric = "euclid")
points(X[sel$model, ], pch = 19, col = 2)

prospectr documentation built on Aug. 31, 2022, 5:05 p.m.