guidedProjections: Guided Projections Data Transformation

Description Usage Arguments Details Value Author(s) References Examples

Description

Computes the orthogonal and score distances and OSD for guided projections.

Usage

1
guidedProjections(data, k=10, osd=c("OD", "SD", "OD*SD")

Arguments

data

A data matrix of n observations and p variables.

k=10

The number of observations used to define the projections.

OSD

Used as similarity measure for the replacement of observations and for the comparison of observations. OSD can be set to "OD" for orthogonal distances, "SD" for score distances and "OD*SD" for the product of orthogonal and score distances.

Details

Computes the orthogonal and score distances and OSD for a series of projections, providing a description of the datastructure. Each projection is corresponds to the projection onto a space spanned by k observations. The observations get sequencially replaced based on a measure of similarity. Guided projections are capeable of revealing group structure and outliers in a dataset. The transformation is especially helpful for the following situations: * inhomogenous groups of observations (e.g. different covariance structure) * groups located in different subspaces

Value

A list-object of class "guidedprojections"" containing the following matrices:

OD

A matrix of orthogonal distances for each observation (columns) to each projection space (rows)

SD

A matrix of score distances for each observation (columns) in each projection space (rows)

OSD

A matrix measures of similarity for each observation (columns) and projection (rows)

Author(s)

Thomas Ortner (thomas.ortner@tuwien.ac.at)

References

Guided projections for analysing the structure of high-dimensional data https://arxiv.org/abs/1702.06790

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
library(lop)
#might take a minute
gp <- guidedProjections(glass$data)
plot(gp, col=glass$group)

#For details on the data set see
#Lemberge, P., De Raedt, I., Janssens, K. H., Wei, F., and
#Van Espen, P. J. (2000). Quantitative analysis of 16–17th
#century archaeological glass vessels using pls regression of
#epxma and μ-xrf data. Journal of Chemometrics, 14(5-6):751–763.

tortnertuwien/lop documentation built on May 30, 2019, 8:27 a.m.