https://github.com/sag2021/kde_sphere
Module for Kernel-Density Estimation (KDE) on sphere
https://github.com/sag2021/kde_sphere
kernel-density-estimation spatial-statistics spherical-geometry
Last synced: about 2 months ago
JSON representation
Module for Kernel-Density Estimation (KDE) on sphere
- Host: GitHub
- URL: https://github.com/sag2021/kde_sphere
- Owner: sag2021
- License: bsd-3-clause
- Created: 2024-10-27T03:15:20.000Z (6 months ago)
- Default Branch: main
- Last Pushed: 2024-11-04T09:15:50.000Z (6 months ago)
- Last Synced: 2025-01-22T12:11:38.903Z (3 months ago)
- Topics: kernel-density-estimation, spatial-statistics, spherical-geometry
- Language: Python
- Homepage:
- Size: 143 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Overview
Module for Kernel-Density Estimation (KDE) on unit sphere. Created for [^1].
Uses a Fisher distribution (i.e. Fisher-Von-Mises with p=3) as kernel function. No bandwidth selection. Calculation can be performed
in chunks to reduce max. memory requirements.
# Coordinates
Coordinates on the sphere must be specified in spherical polar coordinates. The polar angle (theta) must be in the range [0,PI], with
0 at the north pole, and the azimuthal angle (phi) should be in range [0,2*PI]. Both should be in radians.
# Requirements
The base module only requires numpy. To run the unit tests, matplotlib and scipy are required.
# Unit tests
There are two unit tests. The first tests that the Fisher kernel is correctly normalized. The second computes the KDE estimate
for a uniform distribution and compares the MISE to the known MISE: for the uniform distribution on the sphere, the MISE can be computed analytically.
The parameters for the unit tests are set directly in the scripts.
[^1]: Russell M.B., Johnson, C.L., Gilchrist, S.A.: 2024,"Investigating the Spatial Relationship of Shield Volcanism with Coronae on Venus",55th Lunar and Planetary Science Conference,3040,
URL: https://ui.adsabs.harvard.edu/abs/2024LPICo3040.2582R