In my requirement, purchase order documents having multiple pages(varies from 2-6) and only one page is required that contains Purchase order details. The page is not unique like, from 1-6 any pages the PO details will appear. Any suggestions ?
Tried split the document before Digitize Document step and using Intelligent Keyword Classifier.
The current solution that can do this is the Intelligent Keyword Classifier. No need to split the document individually, as the intelligent keyword classifier would return the predicted classes for each page range. So presumably it would tell you which pages contain a Purchase Order. Documentation: Document Understanding - Intelligent Keyword Classifier (uipath.com)
I have one more doubt. If we are using the ML Skills from the AI Center, and we have only one type of document, there is no need to use the classifier activity. But for splitting the docs do we need to use Intelligent keyword classifier?
There’s no need to use the classifier activity if you know that you’re only receiving one type of document. But the situation that you’re facing is that you have a document with multiple document types contained within, where only a subset is a Purchase Order.
Hence you’ll need to use a classifier to go through the document and tell you which pages contain the Purchase Order, and you’ll then need to extract those pages and send them to DU for processing.