Decipher Classification based on text

[HarikrishnaMura]

Hi All,

I am looking to see if I can classify documents using text on the document. For example, i have POs and Invoices that are submitted in a single PDF. Both PO and Invoice look similar. at the top of the document, it says "Purchase Order" for PO and "Invoice" for invoices. Based on this can Decipher recognize if its a PO DFD or Invoice DFD?

[BenLyons]

Hi Harikrishna,

I think it's quite unlikely that Decipher will be able to differentiate these documents if they are as similar as you say. The classification engine is based on general layout and consistent sections of text.

If the fields are the same though, you may only need a single DFD and you could add an extra field to read the Invoice/PO text to confirm. If there are small differences in the required fields, you could include both sets and train the extraction as required. If additional differences are required it would be worth considering the Specific Versions functionality.

Thanks

Ben