Mario,

I played around with your example a bit, substituting more general
model params (not created from swarming). My source code is here:

- https://github.com/rhyolight/online-anomaly-detection/tree/master/mat_src

Here is a video of it plotting:

- http://youtu.be/tw1dmxcI8pU

Doesn't look too bad, does it? This is only after NuPIC has seen 5-6K
rows of data, and it's not flagging every QRS complex as an anomaly.
Keep in mind I'm using a threshold to highlight the bottom chart as
red when the anomaly likelihood goes above 0.9.

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


On Tue, Nov 11, 2014 at 1:04 AM, Mario Tambos <[email protected]> wrote:
> Hi guys,
>
> Thank you for your answers and sorry for the delay.
> Here's a repo I've made with the code:
>
> https://github.com/mtambos/online-anomaly-detection
>
> The code for the OPF is in src/cla. It is bassically a modified version of
> hotgym.
> The data is in src/experiments/ecg1_chfdbchf13
> You can run the experiment with: src/run_ecg1_chfdbchf13.py --do_cla
>
>> Hi Mario,
>>
>> I am working on EEG data and anomaly likelihood and had a look at your
>> info. How do you encode the time stamp and feed it into learning anomaly
>> calcs? I could not see any time stamp encoding.
>
>
> Matt is right, the timestamp is in the first column (marked in the header as
> datetime and with flag "T"), but it's not stricktly needed.
>
>>
>> Mario, any chance we can see your code? You haven't called
>> "disableLearning()" on the model instance, have you? If you can push your
>> code to an online repository, I'll try to run it.
>
>
> Nope, I haven't called disableLearning()
>
>>
>> Have you tried using the AnomalyLiklihood instead of the anomaly score?
>> See
>>
>> https://github.com/numenta/nupic/tree/master/examples/opf/clients/hotgym/anomaly/one_gym#anomaly-likelihood
>> for an example.
>
>
> Yes, I'm using the anomaly_likelihood instead of anomaly_score.
>
> Thanks again,
>
> Mario

Reply via email to