Skip to content

Latest commit

 

History

History
38 lines (27 loc) · 2 KB

README.md

File metadata and controls

38 lines (27 loc) · 2 KB
Documentation Build Status

Kernel Density Estimation

KernelDensityEstimation.jl is a package for calculating univariate (1D) kernel density estimates from vectors of data. Its main features (and limitations) are:

  • Uses a Gaussian kernel for smoothing (truncated at 4σ).
  • Supports closed boundaries.
  • Supports processing weighted samples.
  • Provides higher-order estimators to better capture variation in width and slope of distributions.
  • A more sophisticated bandwidth estimator than the typical Silverman rule.
  • Limited to 1D curves — does not support 2D densities.

This package largely implements the algorithms described by Lewis (2019)1 (and its corresponding Python package, GetDist).

Four-panel plot showing KDEs of simple normal, exponential, beta, and gaussian mixture-model samples

Footnotes

  1. A. Lewis. GetDist: a Python package for analysing Monte Carlo samples (2019), arXiv:1910.13970. ↩