dummy-link

Particles

nonparametric clustering with sequential monte carlo

Readme

Particles

Build Status

codecov.io

This package implements three different Bayesian non-parametric methods for inferring the number of clusters in a mixture model, and cluster assignemnts.

  1. A Gibbs sampler which makes multiple passes over the data (batch).
  2. A particle filter using the resampling scheme described by Fearnhead (2004) (online).
  3. A particle filter using the resampling scheme of Chen and Liu (2000) (online).

These can fit variants of a Dirichlet-process mixture model, with a conjugate prior for the cluster parameters. Currently uni- or multivariate normal distribution components are supported (via a NormalInverseChisq and NormalInverseWishart prior). In principle the code could be generalized to use any of the distributions in ConjugatePriors.jl but this requires some additional methods (for doing online updates of the necessary sufficient statistics and calculating the collapsed/marginal likelihoods).

First Commit

11/30/2017

Last Touched

4 months ago

Commits

165 commits

Used By: