The most / least frequent pattern

Hi guys.
I’m curious about how to detect the most or less??? frequent patterns or sequences seen by TP. It will be good to extract theses things.

1 Like

Just to be clear, we do not use the old temporal pooler anymore at Numenta. I will assume you are talking about the temporal memory algorithm.

We are not currently focusing on temporal sequence pooling at the moment, so I can’t answer your question. But here is a relevant topic:

You should keep track of frequency of spatial patterns or columns and TM cells on your own if you want to keep track of the most common patterns or sequences.

Ok i see. So we can imagine a counter based on hash(cols)…