Main software website: http://bigdata.ices.utexas.edu/software/apm/
MATLAB code that trains the Admixture of Poisson MRFs model (also single Poisson MRF or mixture of Poisson MRFs) using the algorithm described in the following paper:
Capturing Semantically Meaningful Word Dependencies with an Admixture of Poisson MRFs
D. Inouye, P. Ravikumar, I. Dhillon
Neural Information Processing Systems (NIPS), 27, 2014.
For more information on the APM model see the following paper:
Admixture of Poisson MRFs: A Topic Model with Word Dependencies (pdf, slides, poster)
D. Inouye, P. Ravikumar, I. Dhillon.
In International Conference on Machine Learning (ICML), pp. 683–691, June 2014.
- (Optional but highly recommended) Parallel Computing Toolbox for MATLAB (Speedup for using parfor loop is almost linear so code will run signficantly faster with toolbox)