Description:

This source directory contains library routines that provide alternative ways of computing generalised second order Stirling numbers used in working with Pitman-Yor and Dirichlet processes (PYP and DP). Included is library routines for posterior sampling on the discount and concentration parameters of the PYP/DP, and some simple demo scripts.

System used for Pitman-Yor processing with topic models, to allow easy scaling to gigabytes of text. Tested on a few versions of Ubuntu Linux and MacOSX and Windows with gcc. In default configuration has no prerequisities. Allows multi-threaded access on Linux version for use in multi-core contexts.

Matching paper with some of the theory at http://arxiv.org/abs/1007.0296, and a series of papers by Buntine and students.

Changes to previous version:

Moved repository to GitHub, and added thread support to use the main table lookups in multi-threaded code.

BibTeX Entry: Download

Corresponding Paper BibTeX Entry: Download

Supported Operating Systems: Linux, Windows, Mac Os X

Data Formats: Agnostic

Tags: Library, Pitman Yor Process, Stirling Number

Archive: download here

Other available revisons

Version	Changelog	Date
1.8	Moved repository to GitHub, and added thread support to use the main table lookups in multi-threaded code.	April 24, 2014, 09:02:17
1.7	This version, in its default form, doesn't need any other libraries, no GSL. Done to make installation on non-Linux boxes easier.	November 8, 2013, 22:48:40
1.6	New call S_UV() allows faster computation for indicator sampling ala Chen, Du and Buntine ECML-PKDD 2011. Tagging of individual tables when multiple discounts are used in the one model.	September 23, 2013, 05:36:08
1.5	Bug fix for samplea() using getval(). Change to samplea() interface allowing mixed dimensions.	January 4, 2013, 02:00:06
1.4	Replaces "libstb 1.0" release with complete rebuild.	September 28, 2012, 13:41:40

Comments

No one has posted any comments yet. Perhaps you'd like to be the first?

Leave a comment

You must be logged in to post comments.