I am new to exploring the features offered in the latest UiPath Document Understanding framework.
I had few clarifications regarding the training process for Classification and Extraction. Such as:
a) How often should one train a set of invoice before we can see an improvement in the classification accuracy?
b) How often should one train a set of invoice before we can see an improvement in the extraction accuracy? Also, what percentage of improvement is usually seen?
Also, I am curious on how the training process actually works. Is it just an update to the json files set up during development or something else?
It would be great if anyone could help me out with the same.