hdp_mixture

Hierarchical Dirichlet process mixture model

This is a library that implements a Gibbs sampler for a hierarchical Dirichlet process mixture of normals. It was originally developed as a method for this paper: http://www.nature.com/nmeth/journal/v14/n4/full/nmeth.4189.html . Accordingly, some code in the library is specific to nanopore sequencing, but there is a general purpose API as well.

Building hdp_mixture requires gcc version 4.9.1 or higher.

Acknowledgements:

hdp_mixture uses source from RANLIB for random number generation, CuTest for unit testing, and sonLib for basic data structures:

• https://people.sc.fsu.edu/~jburkardt/c_src/ranlib/ranlib.html
• https://github.com/benedictpaten/sonLib
• http://cutest.sourceforge.net/