version

CHMMBOX, version 1.2, Iead Rezek, Oxford University, Feb 2001 Matlab toolbox for max. aposteriori e

What's new in version 1.x:

VERSION 1.1

1. estimates the CHMM in the Maximum likelihood framework.

2. is based on the HMM Toolbox by William Penny 

3. You can initialise the state transition matrices (hmm.P) before
   you do chmmtrain.

4. Includes a weighted EM algorithm for training Gaussian Mixture models where    each data point is weighted by an amount `gamma'.
 
5. A feature allowing you to update only selected observation models during tra   ining (ie observation models for some states stay the same).

6. The following observation models have so far been implemented:
   Gaussian, Gaussian with common covariances, Likelihood, AR.

7. also contains: init_ar.m (and dar.m) necessary for initialising 
   AR vectors in HMM-AR model

8. also contains the files: arembed.m, embed.m and arwls.m 

9. also contains another demo: demar2.m which uses init_ar.m

VERISON 1.2

1. estimates the CHMM in the Maximum aposterior framework.

2: requires parameter priors and contains new functions 
   gaussmd.m, wishart.m dirichlet.m. 
   The following observation models have so far been implemented:
   Gaussian, Gaussian with common covariances, Likelihood. 

3: requires parameter priors and contains new functions 
   gaussmd.m, wishart.m dirichlet.m. 

4. does not contain: init_ar.m (and dar.m) necessary for initialising 
   AR vectors in HMM-AR model