Forecasting multiple(10) steps into the future

Hi, I am new to HTMs.
I have a continuous stream of CPUs utilization data. I want to use HTMs on this streaming data and predict the next 10-20 data points in real time. Can anyone please advise me, which example in the HTM open source code to follow and how can I achieve multistep forecasting.

Many thanks

From the examples folder of the repo:

In model_params.py, go to the bottom and change: 'steps': '5', to the desired number of steps.

If you have data from another machine, it should be easy enough to change line 68 of cpu.py to be an iterator for a source of your choosing.

1 Like

Awesome. Thanks!