Aha, so you might be interested in this:

https://github.com/nupic-community/nupic.critic

There is a video demo: https://www.youtube.com/watch?v=rNbntHRdEmM

If you are trying to do anomaly detection on sound input, the model
params in this project might be a good starting point, in which case
you will not have to swarm.



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


On Mon, Sep 7, 2015 at 8:50 AM, miguel angel rodriguez
<[email protected]> wrote:
> No worries Taylor ;),
>
> Well my input data is Fast Fourier Transformtaions, so basicaly I have a
> list of magnitudes ( floats ) which each element in the list corresponds to
> a certain frequency "range" lets say. So each period of time we have a list
> of maybe 200 elements. Basicaly its a sound photograph each period. How do
> you for example swarm over image data  matrixes? It must be something
> simillar in treatment, isn't it?
>
> example:
> FFT_left_channel = [1.45 0.0002 0.23265 0.15548 23.1326 15.2326 0.96588
> ..... 0.0000.2565 0.14782 ]
> frequencies_left_channel (hz) = [0 5 10 15 20 25 .... 19950 20000]
>
> Any ideas? Sure someone has worked with image ir something similar that uses
> big chuncs of vectorized data
>
>
> 2015-09-07 17:36 GMT+02:00 Matthew Taylor <[email protected]>:
>>
>> Miguel,
>>
>> Sorry for the late response.
>>
>> I don't think that swarming supports the list data type, but I am not
>> sure. What exactly does your input data look like?
>>
>> ---------
>> Matt Taylor
>> OS Community Flag-Bearer
>> Numenta
>>
>>
>> On Thu, Sep 3, 2015 at 2:52 AM, miguel angel rodriguez
>> <[email protected]> wrote:
>> > Hi everyone,
>> >
>> > I was looking for any example of use of lists in swarming, because
>> > aparently
>> > it might not be supported (
>> > https://github.com/numenta/nupic/wiki/NuPIC-Input-Data-File-Format )
>> >
>> > Can anyone confrim me this and if so, how have you gone around this?
>> >
>> > Thanks ;)
>>
>

Reply via email to