We are certainly aware of various Duplicate ID errors and will be working on them: https://github.com/numenta/nupic/issues/2556
If you change your model to an anomaly model, you'll get back anomaly scores in addition to predictions. This gives you an indication how well the model is predicting the data. Follow the Hot Gym Anomaly Tutorial to find out how: https://github.com/numenta/nupic/tree/master/examples/opf/clients/hotgym/anomaly/one_gym Regards, --------- Matt Taylor OS Community Flag-Bearer Numenta On Wed, Sep 9, 2015 at 2:01 AM, Takenori Sato <[email protected]> wrote: > Hi, > > Today, I tried NuPIC 0.3.0 on Ubuntu 14.04.3. > Still I get "Duplicate ID" errors for tests, > but anyway, moved forward to the tutorial. > > I followed the video tutorial of Hot Gym Prediction. > > I can't believe time series data is modeled, > and predicted with such a good fit, > by just running swarm. > > Now I wonder if there is any score available to measure model quality. > My model looks like better than Matt's, but not sure at all. > > Thanks! > Takenori >
