ESS: Expected Simplex Skewness Local Dimension Estimation

Description Usage Arguments Details Value Author(s) References Examples

Description

Local intrinsic dimension estimation with the ESS method

Usage

1
essLocalDimEst(data, ver, d = 1)

Arguments

data

Local data set for which dimension should be estimated.

ver

Possible values: 'a' and 'b'. See Johnsson et al. (2015).

d

For ver = 'a', any value of d is possible, for ver = 'b', only d = 1 is supported.

Details

The ESS method assumes that the data is local, i.e. that it is a neighborhood taken from a larger data set, such that the curvature and the noise within the neighborhood is relatively small. In the ideal case (no noise, no curvature) this is equivalent to the data being uniformly distributed over a hyper ball.

Value

A DimEst object with two slots:

dim.est

The interpolated dimension estimate.

ess

The ESS value produced by the algorithm.

Author(s)

Kerstin Johnsson, Lund University

References

Johnsson, K., Soneson, C., & Fontes, M. (2015). Low Bias Local Intrinsic Dimension Estimation from Expected Simplex Skewness. IEEE Trans. Pattern Anal. Mach. Intell., 37(1), 196-202.

Examples

1
2
data <- hyperBall(100, 4, 15, .05)
essLocalDimEst(data, ver = 'a', d = 1)

intrinsicDimension documentation built on June 7, 2019, 5:02 p.m.