On Thu, Dec 3, 2015 at 11:00 AM, Michael Lam <[email protected]> wrote: > Let's say I am during swarming for prediction steps 1, 5, 10 (based on the > hot gym example). When swarm for prediction steps 1,5,10 is done, is there > a file(s) that tells me the prediction models for all these steps are > generated?
It depends on how you swarm, see: - https://github.com/numenta/nupic/wiki/Running-Swarms#running-models-discovered-by-the-swarm - https://github.com/numenta/nupic/wiki/Running-Swarms#running-a-swarm-programmatically Either way, you end up with model parameters that were the best that the swarm found. > Also, currently, you use the original data file to do swarm and > then run prediction analysis with the same file. For example, I have a file > with data from 11 am to 12 pm, with data points per minute. That's not enough data for a swarm. There are 1440 minutes in a day, so I suggest you pass in 3 days worth of data (4320 points) into the swarm. That gives it a chance to identify whether "time of day" is a relevant factor when deciding the model parameters. > I use that file > to generate swarm model, and then run prediction analysis with the same file > in my run.py, which will stop at 12 pm even if I am using the 10 step ahead > prediction model. If I want to predict the values from 12:01 to 12:10, I > will need to create a file with those additional time slots to be used by > run.py, ie 12:01, 12:02, 12:03 ...12:10. So do I put 0 as the value for > those additional time slots? Or am I missing something here? Thanks. You can send as much data as you want into a swarm (but use at least 3000 points IMO). Once you have your model params, you don't need to send the same data file into the model as you used to swarm. Swarming can be done over some sample data set and the real data sent into a running model. The idea is to swarm as a pre-processing step, then instantiate a model using the params a swarm creates, passing live data into the model at that point. Regards, --------- Matt Taylor OS Community Flag-Bearer Numenta
