search

markovmodel / PyEMMA

🚂 Python API for Emma's Markov Model Algorithms 🚂
http://pyemma.org
GNU Lesser General Public License v3.0
310 stars 119 forks source link

PCCA for non-reversible Markov Chains #1455

Closed Marius1311 closed 4 years ago

Marius1311 commented 4 years ago

Hi all, I was trying to compute a PCCA for a given transition matrix but got an error because the Markov Chain described by the transition matrix was not reversible, see https://github.com/markovmodel/PyEMMA/blob/c50da3fd1f517b07a45a96d8aea1539bbd44cbd5/pyemma/msm/models/msm.py#L936

I assume that this is the expected behaviour of the code right now. However, in the description of the method, you cite Roeblitz and Weber (2013), who write in their abstract:

We therefore demonstrate in this paper that PCCA+ always delivers an optimal fuzzy clustering for nearly uncoupled, not necessarily reversible, Markov chains with transition states.

So I was wondering whether the code presented here is actually an implementation of their ideas and thereby PCCA++, or whether it's really PCCA+, which is only expected to run for reversible MCs.

stale[bot] commented 4 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.