scalemap 0.1.1

A modification of umap that incorporates local scale into the embedding.

ScaleMAP

ScaleMAP is a modification of UMAP that incorporates local scale into the embedding.

This package is currently a quick release. It is derived from umap-learn and keeps the same scikit-learn-style estimator interface.

Installation

pip install scalemap

Usage

import scaleMAP

mapper = scaleMAP.UMAP(scalemap=True, scale_lambda=1.0)
embedding = mapper.fit_transform(data)

The distribution name on PyPI is scalemap; the import package is scaleMAP. If you don't set scalemap=True, it will behave just like UMAP. scale_lambda=1.0 sets the scaleMAP behavior to be fully on. You can interpolate between fully on and fully off by setting values between 0.0 and 1.0.

Attribution

ScaleMAP is derived from umap-learn by Leland McInnes and contributors. The upstream project is available at:

https://github.com/lmcinnes/umap

The upstream UMAP package is BSD 3-Clause licensed. ScaleMAP is distributed under the same license. See LICENSE and NOTICE.md.

If you use this package in scientific work, cite the original UMAP work as appropriate:

@article{mcinnes2018umap-software,
  title={UMAP: Uniform Manifold Approximation and Projection},
  author={McInnes, Leland and Healy, John and Saul, Nathaniel and Grossberger, Lukas},
  journal={The Journal of Open Source Software},
  volume={3},
  number={29},
  pages={861},
  year={2018}
}