Conditional Random Field (CRF) Toolbox for Matlab

Conditional Random Field (CRF) Toolbox for Matlab

1D chains

2D lattices

Written by Kevin Murphy, Mark Schmidt (2006).
Last updated: 23 May 2006.

This is the code that was used to create the 2D results in this paper
Accelerated Training of Conditional Random Fields with Stochastic Meta-Descent
S Vishwanathan, N. Schraudolph, M. Schmidt, K. Murphy
ICML'06 (Intl Conf on Machine Learning).
Note: the published version has some small errors in the experimental results, so the figures will not exactly match the code below. Click here) for a revised version. Now, figures 4 and 6 give can be reproduced using the online code available below.

General graphs

Written by Kevin Murphy, 2004.
This supports 1D chains, 2D lattices and general graphs.
This code is no longer supported.
The 2D lattice code (above) is better engineered and can be extended to handle general graphs quite easily.

Other resources for CRFs