ClusterPy

Bayesian Hidden Markov Model with Gibbs sampling

Requirements

Python 2.7.3+

Usage

./cluster.py <input file> [output file] <number of labels> <number of iterations> <alpha> <beta>

Arguments

<input file> - File with sequences, one per line
[output file] - Output file
<number of labels> - Possible labels
<number of iterations> - Number of sampling iterations
<alpha> - Hyperparameter for transitions
<beta> - Hyperparameter for emissions

Content

• bhmm.py - Gibbs sampling
• cluster.py - Entry point
• utility.py - Utility methods
• corpus.txt - toy corpus

License

MIT