geokde-t2 0.1.0

Geospatial Kernel Density Estimation (KDE) on Toroidal Manifolds (Standalone Module)

geokde-t2

Geospatial Kernel Density Estimation (KDE) on Toroidal Manifolds (Standalone Module)

This is a standalone module of the axiom-t2 geometry-aware machine learning toolkit for toroidal manifolds. It provides standalone packaging for applications requiring a lightweight dependency footprint.

Installation

pip install geokde-t2

Quick Start

from geokde import GeoKDE, flat_kde
import numpy as np

# Fit GeoKDE on geospatial data in radians
X = np.random.uniform(-np.pi, np.pi, (100, 2))
kde = GeoKDE()
kde.fit(X)
scores = kde.score_samples(X[:5])
print("Log-likelihoods:", scores)

Reference

For the underlying mathematics and details on toroidal manifolds, please refer to the main toolkit paper: axiom_t2_toolkit_paper.pdf.