Martin K. Scherer

Agreed upon the fear on code obfuscation in this spot. The sampling should be implemented as partial_fit in BayesianMSM to be clear. However the user would then have to access...

I think returning multiple models/estimators upon a parameter is very confusing and also breaks the API. Why don't we extend ITS to do handle TICA too?

@MamtaMohan you can now use msmtools-1.2 for rate matrix estimation

the binaries the released version are also available ;) just do a conda update msmtools

On 11/16/2016 03:46 AM, Frank Noe wrote: > Hi, > > you may need to update msmtools: > conda install msmtools > > then your import statement should work as...

That is a nice improvement. We just need to substitute the metric calculations by ours and should be good to go. MIT license is permissive, right? So we just leave...

Do you need to have the same discretization for both matrices or does the method recover for slightly different state spaces?

The order of the intermediate sets is random because the coarse graining uses Python set operations to calculate intersections etc. Sets are unordered data structures, this is why we will...

can you dump this object and upload is somewhere?

tica_obj.save('tica_debug_koopman.pyemma')