Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
  • #2455
    Omar Stradella

    It looks like the only way to test models is by actually scanning samples. I would rather have the ability to choose a training set and a test set from the set of all records. That would make tweaking the model much faster and easier.


    Dear Omar,

    We responded via e-mail as well.

    Defining a test set is in our feature list, however notice that as long as the collection is big enough, the current cross validation process should be fine and provide accurate results.

    Would you prefer to download the raw spectrum and use your own analysis tools? 

    This feature will also be available in the future by upgrading to SCiO’s Researcher kit. 



    The ConsumerPhysics Team

Viewing 2 posts - 1 through 2 (of 2 total)
  • You must be logged in to reply to this topic.