Reduces the number of random points in a hypervolume

Share:

Description

Many hypervolume algorithms have computational complexities that scale with the number of random points used to characterize a hypervolume (@RandomUniformPointsThresholded). This value can be reduced to improve runtimes at the cost of lower resolution.

Usage

1
hypervolume_thin(hv, factor = NULL, npoints = NULL)

Arguments

hv

An object of class Hypervolume

factor

A number in (0,1) describing the fraction of random points to keep.

npoints

A number describing the number random points to keep.

Details

Either factor or npoints (but not both) must be specified.

Value

A Hypervolume object

Examples

1
2
3
4
5
data(iris)
hv1 = hypervolume(subset(iris, Species=="setosa")[,1:4],bandwidth=0.2)

# downsample to 1000 random points
hv1_thinned = hypervolume_thin(hv1, npoints=1000)