Edit on GitHub

Datasets

The ManifoldLearning package provides an implementation of synthetic test datasets:

ManifoldLearning.swiss_roll — Function
swiss_roll(n::Int, noise::Real=0.03, segments=1)

Generate a swiss roll dataset of n points with point coordinate noise variance, and partitioned on number of segments.

source
ManifoldLearning.spirals — Function
spirals(n::Int, noise::Real=0.03, segments=1)

Generate a spirals dataset of n points with point coordinate noise variance.

source
ManifoldLearning.scurve — Function
scurve(n::Int, noise::Real=0.03, segments=1)

Generate an S curve dataset of n points with point coordinate noise variance.

source