Clustering Data using Mixtures of von Mises-Fisher Distributions

This web page offers a C++ software for clustering directional data. The data is modeled as a mixture of K von Mises-Fisher distributions, whose parameters are estimated using Expectation Maximization (EM). The code contains two versions of EM, the first one yields a hard or partitional clustering and the second version is the standard soft-clustering EM algorithm.

The original MoVMF code was written by Arindam Banerjee and Suvrit Sra.

Downloads

Documentation

Citation

Clustering on the Unit Hypersphere using von Mises-Fisher Distributions Arindam Banerjee, Inderjit S. Dhillon, Joydeep Ghosh, Suvrit Sra
Journal of Machine Learning Research    6(Sep):1345--1382, 2005.

      A forerunner to the above paper Generative Model-based Clustering of Directional Data Arindam Banerjee, Inderjit S. Dhillon, Joydeep Ghosh, Suvrit Sra.
      ACM SIGKDD, Washington DC. August, 2003

BibTeX entries

Last modified: Wed Nov 7 03:14:02 CST 2007