hmm-kit 0.3.0

Python toolkit for unsupervised learning of sequences of observations using HMM

hmm_kit

Simple Hidden Markov Models library.

Supports discrete/continuous emissions of one/multiple dimensions.

Install

python setup.py install

Usage

See examples in tests/hmm_test.py

Standalone script

To run it as a standalone script call simple_hmm.py

• Training model simple_hmm.py train_model train_data.csv --kernel gaussian --states 3 --model_name my_model
• Decode - simple_hmm.py decode train_data.csv my_model decode.csv
• Posterior - simple_hmm.py posterior train_data.csv my_model posterior.csv