DPMM.jl

parallel gibs sampler for dirichlet process mixture models

Date: 9 July 2019

This project is a research work on distributed inference for Dirichlet Process Mixture Models on Julia by me with the supervision of John W. Fischer III. I implemented distributed Split-Merge Gibbs sampler in Julia which is both modular and as efficient as C++ implementation (see benchmark results). Here are some resulting clusters from NYTimes articles dataset.