See also and classify - can parts of text be excluded?

I quite like the “See also and Classify” function as part of my patent portfolio set up. Having set up a group structure of technology domains prior, I can now use “SaaC” to determine where a new patent or innovation proposal probably belongs and how strongly it is associated with the existing text.

I cannot tell you enough how handy this is.

However, patent text tends to come with a bunch of boiler plate I do not wish to be considered by SaaC - it might throw of its targeted classification. If I could highlight the important bits or remove the unimportant bits for SaaC that would be great. Is there a way to do this? Note that I still need the boiler plate text ones I store it in DT, I just don’t want it to be part of SaaC.

If the undesired text is included in the same document as the desired text (for “Saac” purposes), the answer is no – you cannot mark parts of a document to exclude. However, you can exclude who documents. Look at Tools > Show Info > Exclude …. A possible solution then is to split PDFs into two (or more) portions – the “boilerplate” and the rest. Use Exclude… on the “boilerplate” document.

If I were doing this, I would also add a label or a distinctive tag to the “excluded” document to indicate its excluded status.

Yes I thought about doing what you suggest i.e. split the document. But I think it is a bit too manual. Besides I think keeping two parts of what is essentially 1 document around - associated with each other, is creating problems I think.

Bummer. I know DT is not a one purpose tool and my request/question is very specific. I had just hoped DT could be even better for my purposes. It is such a great system for patent research…

I find this thread interesting because auto-classification improvements are one of two areas I painfully wish to see improvement in DTP (the other being manual classification improvements). DTP’s classification features are so unique and have so much potential.

Personally, I would like the option to expand the scope of See Also & Classify (document name, e-mail subject, sender, etc… not just body). But no doubt being able to restrict would also be interesting and useful in ways I’ve probably not even considered…

Take a look at See Related Text as a possibility of “restricting” See Also in cases where documents may contain multiple topics. In that case, select just the text of the desired topic, Control-click (right click) on it and choose the contextual menu option See Related Text. This will probably better focus the list of suggestions, compared to See Also.

Tip: When I want to explore one or more suggestions made by DEVONthink, I Control-click on a suggestion to open it as a new tab, rather than replace the display of the document I’m reading.

Had no idea this was available - thank you!

Indeed that is helpful Bill ! I had missed that completely.