computeSampling: Sampling raw data matrix

In RclusTool: Graphical Toolbox for Clustering and Classification of Data Frames

Sampling raw data matrix

Description

computes sampling on raw data matrix to reduce the number of observations, with generalization step.

Usage

computeSampling(
  x,
  label = NULL,
  K = 0,
  toKeep = NULL,
  sampling.size.max = 3000,
  K.max = 20,
  kmeans.variance.min = 0.95
)

Arguments

x	matrix of raw data (point by line).
label	vector of (named) labels.
K	number of clusters. If K=0 (default), this number is automatically computed thanks to the Elbow method.
toKeep	vector of row.names to keep in the sample (for constrained algorithms).
sampling.size.max	maximal number of observations to keep in the sample.
K.max	maximal number of clusters (K.Max=20 by default).
kmeans.variance.min	elbow method cumulative explained variance > criteria to stop K-search.

Details

computeSampling computes sampling on raw data matrix to reduce the number of observations, with generalization step.

Value

The function returns a list containing:

selection.ids	vector of selected row.names in the sample.
selection.labs	vector of selected labels in the sample.
matching	character specifying the matching for all observations and used for generalization of the clustering result.
size.max	maximal number of observations kept in the sample.
K	number of clusters.

Examples

dat <- rbind(matrix(rnorm(100, mean = 0, sd = 0.3), ncol = 2), 
             matrix(rnorm(100, mean = 2, sd = 0.3), ncol = 2), 
             matrix(rnorm(100, mean = 4, sd = 0.3), ncol = 2))
tf <- tempfile()
write.table(dat, tf, sep=",", dec=".")
x <- importSample(file.features=tf)

res.sampling <- computeSampling(x$features$initial$x)

RclusTool documentation built on May 29, 2024, 5:23 a.m.