What algorithm is used by prodigy's text classification recipe?

I want to use prodigy for some academic research as part of my PhD and need to explain the algorithm used. I've had a look at the FAQ and text classification pages but can't see where I can find details of the approach used. I don't need the code or fine details - a high-level overview would do ideally somewhere I can provide a link to in my references.
Is this information available in the documentation somewhere or is this commercially sensitive and if I need to explain the algorithm then I'll need to use something other than prodigy's recipe?

Hi @nicfox ,

By default, textcat uses a stack ensemble of linear BOW (bag-of-words) plus a neural network. You can check the architecture in the documentation. Note that this can change depending on how you configured it, so it might be best to look at the overall list instead.

1 Like

thanks @ljvmiranda921 :hugs: