What is the advantage of textcat.teach over .correct?
textcat.correct also label using similarity like texcat.teach?
thanks
Both recipes are indeed similar. The difference are explained in more detail in the textcat.correct docs and the textcat.teach docs but the main difference is that in the base behaviour the teach recipe has a model in the loop that updates as you give it labels while the correct recipe does not.
That said, you can configure textcat.correct to also allow the model to update and you can configure a --threshold to tune how examples are pre-selected for annotation. The textcat.teach recipe, on the other, allows for a --patterns parameter. This allows you to supply a path to match patterns file to filter out examples containing terms and phrases. The matched spans are also highlighted when labelling.
Sorry , but on the doc, say
"Create training data for an existing trained text classification model by correcting the model’s suggestions"
and on the example, pre-ckeck the labels over a threshold,
That is more efficient than teach because can annotate and update vs teach that show the same item multiple times, even after to check the right suggestion
That my confusion
It's perhaps good to also mention that the user interfaces are also different. The teach-variant gives you a binary interface, which only allows you to say yes/no to a model's prediction. The correct-variant is indeed a manual interface, which allows you to label multiple entities but will require you to use your mouse.