Although I do not know, I would also assume that if the sequence was
reset due to input data tripping reset logic, the
`result.sensorInput.sequenceReset` would be `1`.

This also supports that idea:
- 
https://github.com/numenta/nupic/blob/master/nupic/frameworks/opf/opfbasicenvironment.py#L142-L144
- 
https://github.com/numenta/nupic/blob/master/nupic/frameworks/opf/opfutils.py#L189-L191

I will have to ask someone else to answer this question.

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


On Wed, Sep 16, 2015 at 8:55 AM, Kearney, Conor
<[email protected]> wrote:
> I watched the hotgym example videos.  From somewhere (not sure where), I got 
> this line of code:
>
> datasetPath = resource_filename("nupic.datafiles", "extra/hotgym/hotgym.csv")
>
> When running the example with this data, would you expect 
> result.sensorInput.sequenceReset to equal 1 when gym changes?
>
> -Conor
>
> -----Original Message-----
> From: nupic [mailto:[email protected]] On Behalf Of Matthew 
> Taylor
> Sent: Wednesday, September 16, 2015 4:26 PM
> To: Kearney, Conor
> Subject: Re: sequence reset
>
> That data file is archived, and I'm not sure it is actually used anywhere in 
> our examples. It was the original source of our "hot gym"
> data. But to answer your question, the "S" field indicator means that this 
> field indicates a sequence. It means when this field changes, a new sequence 
> is being indicated. This helps in exactly this case where the field is a 
> label for the gym the data is coming from.
>
> Out of curiosity, how did you come across this file? Did you see our other 
> hot gym tutorials?
> https://github.com/numenta/nupic/wiki/Using-NuPIC#tutorials
>
> ---------
> Matt Taylor
> OS Community Flag-Bearer
> Numenta
>
>
> On Wed, Sep 16, 2015 at 7:57 AM, Kearney, Conor <[email protected]> 
> wrote:
>> From the https://github.com/numenta/nupic.git repo:
>>
>> nupic/datafiles/extra/hotgym/hotgym.csv
>>
>> -Conor
>>
>> -----Original Message-----
>> From: nupic [mailto:[email protected]] On Behalf Of
>> Matthew Taylor
>> Sent: Wednesday, September 16, 2015 3:40 PM
>> To: Kearney, Conor
>> Subject: Re: sequence reset
>>
>> Conor,
>>
>> Exactly which hotgym example are you running? The tutorials I've put 
>> together don't use resets or sequences in the input CSV data. I think you 
>> found some lower-level examples of swarming. Please tell us exactly what CSV 
>> files in the codebase you are running as input data.
>>
>> Thanks,
>> ---------
>> Matt Taylor
>> OS Community Flag-Bearer
>> Numenta
>>
>>
>> On Wed, Sep 16, 2015 at 7:35 AM, Kearney, Conor <[email protected]> 
>> wrote:
>>> Hello,
>>>
>>>
>>>
>>> I am running the hotgym example and have a question about the
>>> sequence reset mechanism. I note in the csv, the column 'gym' has an 'S' 
>>> tag for sequence.
>>> When running the model, I check the value of sequenceReset from
>>> sequenceReset = result.sensorInput.sequenceReset.
>>>
>>>
>>>
>>> This value is never 1, despite the gym name changing 5 times in the csv.
>>>
>>>
>>>
>>> Am I checking in the correct place for the sequenceReset flag?
>>>
>>> Does sequenceReset need to be explicitly stated somewhere in the json
>>> for swarming or model_params for prediction?
>>>
>>>
>>>
>>> I assumed if the flag is marked as 'S' in the data csv then the reset
>>> would be in effect for swarming and prediction.
>>>
>>>
>>>
>>> Thanks,
>>>
>>> Conor Kearney
>>>
>>>
>>>
>>> Conor Kearney
>>> Data Scientist
>>> Product Operations - QR&A
>>> Allstate Northern Ireland
>>> 9 Lanyon Place
>>> Belfast
>>> Phone: 847-326-8495
>>> Allstate Ext: 68495
>>> UK Phone: 02890678495
>>> Email: [email protected]
>>>
>>>
>>>
>>> The information contained in this message may be privileged and
>>> confidential and protected from disclosure.
>>> If the reader of this message is not the intended recipient, or an
>>> employee or agent responsible for delivering this message to the
>>> intended recipient, you are hereby notified that any dissemination,
>>> distribution or copying of this communication is strictly prohibited.
>>>
>>> If you have received this communication in error, please notify us
>>> immediately by replying to the message and deleting it from your computer.
>>>
>>>
>>
>

Reply via email to