Hi Matt,

We've managed to get it to output both timestep predictions (+1 & +5) to the 
output file and to plot on matplotlib by editing the nupic_output.py
Thanks for your advice.

Cavan


-----Original Message-----
From: nupic on behalf of Matthew Taylor
Sent: Tue 8/5/2014 3:32 PM
To: NuPIC general mailing list.
Subject: Re: [nupic-discuss] Multiple Predictions in Hotgym Tutorial [NP]
 
Whoops, I replied to your first email without reading the second. So you've
already figured out how to plot a different prediction step, but to plot
both in matplotlib, you'll have to do some legwork with matplotlib in the
nupic_output.py files for the respective examples.

I've thought about creating some type of NuPIC output plotter that will
handle lots of situations and accept the entire model result object, and
adding it to the OPF. Seems like that would be useful to you?

---------
Matt Taylor
OS Community Flag-Bearer
Numenta


On Tue, Aug 5, 2014 at 2:50 AM, Naomi Basu <[email protected]> wrote:

>  Classification: NPL Management Ltd - Public
>
> I forgot to mention we also added a "prediction2" variable to the "run.py"
> file as shown:
>
>
>
> prediction = result.inferences["multiStepBestPredictions"][1]
>
> prediction2 = result.inferences["multiStepBestPredictions"][5]
>
> output.write([timestamp], [consumption], [prediction], [prediction2])
>
>
>
> We then ran "run.py", but still no +5  predictions column or graph. We are
> however able to generate +1 and +5 predictions one at a time.
>
>
>
> Thanks,
>
> Naomi
>
>
>
>
>
> *From:* nupic [mailto:[email protected]] *On Behalf Of *Naomi
> Basu
> *Sent:* 05 August 2014 10:06
> *To:* [email protected]
> *Subject:* [nupic-discuss] Multiple Predictions in Hotgym Tutorial [NP]
>
>
>
> Classification: NPL Management Ltd - Public
>
> Hi Matt,
>
>
>
> I have been working with Cavan and we have successfully run swarms over
> the Hotgym and Sine Wave examples, with  +1 step predictions.
>
> Is it possible to get +1 and +5 predictions at the same time, and maybe
> graph them both using matplotlib?
>
>
>
> We tried making the value of "predictionSteps" in the swarm_description
> file [1, 5], however, after running the swarm again, we still get just one
> prediction column in the "rec-center-hourly_out.csv" file.
>
>
>
> Thanks for your help,
>
> Naomi
>   ------------------------------
>
> If you have received this message in error, please notify us and remove it
> from your system.
> NPL Management Ltd cannot guarantee that the e-mail or any attachments are
> free from viruses.
>
> NPL Management Ltd is a company registered in England and Wales, number:
> 2937881
> Registered office: Serco House | 16 Bartley Wood Business Park | Hook,
> Hampshire | UK | RG27 9UY
>  ------------------------------
>
> [image: cid:[email protected]] <http://www.npl.co.uk/>
>
> STAY CONNECTED
> [image: cid:[email protected]] <http://www.twitter.com/npl>[image:
> cid:[email protected]] <http://www.facebook.com/npldigital>[image:
> cid:[email protected]] <http://www.youtube.com/npldigital>[image:
> cid:[email protected]]
> <http://itunes.apple.com/us/podcast/national-physical-laboratory/id295076289>[image:
> cid:[email protected]]
> <https://ktn.innovateuk.org/web/measurement-network>
>
>
>
> _______________________________________________
> nupic mailing list
> [email protected]
> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>
> _______________________________________________
> nupic mailing list
> [email protected]
> http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org
>
>




--
If you have received this message in error, please notify us and remove it from 
your system.
NPL Management Ltd cannot guarantee that the e-mail or any attachments are free 
from viruses.
 
NPL Management Ltd is a company registered in England and Wales, number: 2937881
Registered office: Serco House | 16 Bartley Wood Business Park | Hook, 
Hampshire | UK | RG27 9UY

_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org

Reply via email to