SAP Document Information Extraction result training

[former_member802816]

Hi,

We are using SAP document information extraction service to extract the PDF, but some value is not correct, and I already enabled dataFeedbackCollection, but seems no check box enabled in UI

In our case, the dockCode is "5Z", but extract result is "57", how to make the value correct by enable dataFeedbackCollection?

Best Regards,

Gary

[tobias_weller · Accepted solution]

Hi Gary,

Thanks for sharing the document with us for retraining.

The issue you observe is however caused by the underlying character recognition technology and not by the AI technology used for extracting the entities from the document. However, the retraining we offer here is currently limited on the extracting part and not the character recognition.

It would be best if you can create a support ticket on component CA-ML-BDP-OCR and share the document through the support ticket with the responsible team. They can also give more insights if and how quickly this can be improved.

Best regards,
Tobias

[former_member802816 · Accepted solution]

Hi Weller,

Thanks a lot.

Already submit an incident!

Best Regards,

Gary