No project description provided
Project description
FindTheGap
This package provides tools for geometric data analysis, targeted at finding gaps and valleys in data distribution. It provides a (twice-differentiable) density estimator (Quartic Kernel Density Estimator) relying on pytorch for auto-differentaition, and methods to estimate critical points in the density as well as various statistics to identify and trace `gaps' and valleys in the distribution.
This package can be installed through pip (https://pypi.org/project/findthegap/):
pip install findthegap
Dependencies:
-
numpy >= 1.19.5
-
torch >= 1.10.1
-
scipy >= 1.5.4
Notebook requirements: sklearn, matplotlib
The folder 'examples' contains a notebook showcasing how to use those tools on 2D data (available in the folder data).
Disclaimer: this code is work in progress and might go through some changes especially for higher (>2!) dimension...
Contributors: Gabriella Contardo (CCA at Simons Foundation), David W. Hogg(CCA/NYU/MPIA), Jason S.A. Hunt (CCA)
You can find more information about the methods in the paper "The emptiness inside: Finding gaps, valleys, and lacunae with geometric data analysis" https://arxiv.org/abs/2201.10674
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for findthegap-0.0.10-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2e4e65249551257b6ab028821db58d1a1ddd4e079ed0972c9a4dde0bf50e0ee3 |
|
MD5 | 5a2487edf91d799f199d165ff7469484 |
|
BLAKE2b-256 | 535fa25be311564e69d18db2d0a51c1d65f705d4568239e5e3a10cdc4b194f3f |