hmm is a Python module designed to work with Hidden Markov Models. The usual algorithms (Viterbi, Baum-Welsh) are implemented using Numeric Python.
|Tags||Scientific/Engineering Artificial Intelligence Mathematics Software Development Libraries Python Modules|
|Operating Systems||OS Independent|
Release Notes: This release adds minor bugfixes.
Release Notes: An important bug preventing convergence when learning long inputs was removed. Key parts of the Baum Welsh algorithm were rewritten in C for performance (x150 speedup).
Release Notes: This release adds two learning algorithms, supervised learning and unsupervised learning from several short examples.
Release Notes: Some errors which prevented the viterbi algorithm from working in some cases was fixed.
No changes have been submitted for this release.