I extracted some fields from an invoice and then after that, I used (Present validation station) so that I can improve the extraction.
There was an issue with one field so I corrected it during validation but when I run the program again the same issue was still there in Validation Station.
Is this normal behavior ??
If yes, then what’s the purpose of validation??
Yes, this is normal behavior.
Validation Station is only for correction. Output of this activity have corrected data with 100% confidence - you can use it in further processing. But until you use corrected data in training scope to improve data extraction model, the result of the next run will be the same.
What do you mean by “use corrected data in training scope to improve data extraction model”?
can you elaborate a bit
Currently I am using invoices in PDF to extract data from
Validation station does not improve your extractors. For this purpose there is train extractors scope - https://docs.uipath.com/activities/docs/train-extractors-scope. Note that not every extractor can be trained. I recommend you to see document understanding process template - Document Understanding Process: Studio Template and investigate how extractor training is implemented there.