distortions: distortions

In semisupKernelPCA: Kernel PCA projection, and semi-supervised variant.

Description

the same objects are represented by two different datasets: this function computes the pairwise distances in both datasets, and the distortions induced by these distance matrices (see Aupetit, 2007)

Usage

distortions(origdata, projdata)

Arguments

origdata	n x d original dataset
projdata	n x q projected/transformed dataset

Value

list object, with components: - compress : vector of n compressive distortion values - stretch : vector of n stretching distortion values

Author(s)

Pierrick Bruneau

References

Aupetit, M. (2007) Visualizing distortions and recovering topology in continuous projection techniques. Neurocomputing. Pages 1304-1330.

See Also

computeKernel computeProjectionFromKernel

Examples

data(iris)
irisdat <- as.matrix(cbind(iris[,1:4], rep(NA, 150)))
irisdat[sample(1:150,4),5] <- 1:4
proj <- computeProjectionFromKernel(computeKernel(irisdat))
dists <- distortions(irisdat[,1:4], proj)

Example output

Loading required package: datautils
Loading required package: deldir
deldir 0.1-16
Loading required package: gplots

Attaching package: 'gplots'

The following object is masked from 'package:stats':

    lowess

Loading required package: irlba
Loading required package: Matrix

semisupKernelPCA documentation built on May 29, 2017, 8:59 p.m.