Hi @subutai

Thank you for your quick answer. I tried taking inverse log of the sum of logs of M components and subtracting that from 1 and it seems it is working.

I’ll post an update on it works on my data once it is finalized.

However, I have additional problem now, I need to save the model after observation of a huge data, for sake of performance ( it takes so long to test the huge data which is common in several data files )

I tried using WriteToCheckPoint() method in nupic/src/nupic/frameworks/opf/model.py, However I get error global name “HTMPredictionModelProto” is not defined. Do you have any suggestion how to solve this problem? or is there other way to save the model and re-use it later?

Thanks,