Hi Ed, Rodrigo, and Joe,

Thank you for your replies and tips! Ed, your question is similar to what I was thinking, but my focus was more on research data and non-traditional academic resources. I remembered that citation styles are increasingly complex, and the reported publication types in article databases are not always accurate. If we can take a photo of a plant and get its name right, why not with citation styles? I was curious if someone has already figured this out so I could perhaps use their solution.😊 

AI classifier and Flair NLP -- thank you!

Best,
Sarah G. Park

-----Original Message-----
From: EDWIN VINCENT SPERR <[log in to unmask]> 
Sent: Monday, September 30, 2024 8:23 AM
To: Code for Libraries <[log in to unmask]>
Cc: Park, Sarah <[log in to unmask]>
Subject: RE: CODE4LIB Digest - 28 Sep 2024 to 29 Sep 2024 (#2024-202)

Sara -- 

Were y'all looking to do some sort of broad bibliometric study along the lines of "Particle physics has 20% fewer 'Review' articles than structural engineering"?

If so, I'd be temped to pull out a random sample from each discipline and classify them manually...

(Of course, that's easy for me to say as the person who won't be trawling through hundreds of records)

Otherwise, as Rodrigo suggests, you're gonna be building a classifier...


Ed Sperr, MLIS
Systems and Discovery Librarian
[log in to unmask]
University of Georgia Libraries, Main Library Athens, GA 30602-1641

Date:    Sun, 29 Sep 2024 22:25:58 +0000
From:    "Park, Sarah" <[log in to unmask]>
Subject: identifying publication types from citations

Hi,

I am looking for a tool or method that can help us identify publication types from citations/references using scripts or AI-based tools. My colleague and I are interested in citation analysis to determine the types of sources used in a discipline, for example, journal articles, review articles, magazine articles, book chapters, books, websites, government documents (Gov Docs), and NGO documents.

One possible method I got so far was using article database APIs, like Scopus, to identify document types, but Scopus seems to track some types but not all. I also heard that a model can be trained using ChatGPT or other generative AI, but I haven't heard how effective it can be.

Any thoughts or suggestions that could lead to a possible solution would be greatly appreciated!

Best,


Sarah G. Park, she/her
Mathematics and Computational Sciences Librarian Head, Mathematics Library Assistant Professor University of Illinois at Urbana-Champaign [log in to unmask]<mailto:[log in to unmask]>