plot.ppci_projection_solution

a solution arising from one of the functions mcdr, mddr and ncutdr.

(optional) a vector of class labels. if provided then points in different classes are plotted in different colours.

labels

(optional) if omitted then the first two dimensions are shown. If an integer &gt; 1 then all pairs from the first 1:pairs dimensions are shown.

pairs

(optional) logical. If TRUE then an additional principal component rotation is applied. This can make more clusters visible within a low dimensional projection.

(optional) if ommitted then points in scatterplots are
 shown as solid. If set to a value in (0, 1) then points are shown with transparency. Lower values correspond with a greater degree of transparency.

transparency

additional graphical parameters. Currently none implemented.

Provides a visualisation of a data set using a low dimensional projection obtained using projection pursuit based on cluster based projection indices.

internal

Implements recently developed projection
pursuit algorithms for finding optimal linear cluster
separators. The clustering algorithms use optimal
hyperplane separators based on minimum density, Pavlidis et. al (2016) <http://jmlr.org/papers/volume17/15-307/15-307.pdf>;
minimum normalised cut, Hofmeyr (2017) <doi:10.1109/TPAMI.2016.2609929>;
and maximum variance ratio clusterability, Hofmeyr and Pavlidis (2015) <doi:10.1109/SSCI.2015.116>.

plot.ppci_projection_solution: Visualise a Data Set Projected from Projection Pursuit

Description

Usage

Arguments

Value

See Also