SDR Classifier prediction confidence is NaN

[rhyolight]

As reported by Phil at https://discourse.numenta.org/t/sdrclassifier-returning-confidence-of-nan/5972:

I’ve created various models where after the model has run for some time the classifier confidence value starts returning NaN.

As a sanity check, I’ve gone back to the standard hotgym example (http://nupic.docs.numenta.org/stable/quick-start/network.html) and found that it too suffers from this problem.
For data, I am using the rec-center-hourly.csv file (https://github.com/numenta/nupic/tree/master/examples/opf/clients/hotgym/prediction/one_gym), but I am replicating/extending the data to have 30000 time steps (rather than the default 4394 times points).
Nothing else is being changed from the default hotgym code/model.
Specifically, at about the 19000th data point, the 5-step ahead confidence starts returning NaN.

[PhilG4tgc]

I have attached a code package hotgym_test_code.zip that reproduces the problem.

Executing hot_gym.py will run the model for 18819 steps.
At step 18809, and then from step 18815 onwards, the 5-Step ahead confidence interval is NaN.

The model is identical to the standard hotgym Network API example except for also displaying the time step index.

The installed packages are
numpy==1.12.1
nupic==1.0.5
nupic.bindings==1.0.6
PyYAML==3.10

It is being executed on a Windows 10 box.

[rhyolight]

I replicated the bug. I think it is a bug in the C++ SDR Classifier. The Python implementations did not have this issue. More details:

• https://discourse.numenta.org/t/sdrclassifier-returning-confidence-of-nan/5972
• https://www.twitch.tv/videos/422636647

[suzusuzu]

@rhyolight
please show this
numenta/nupic.core-legacy#1453