One of the products I was working on, Classify, is meant to help non-subject experts within publishing organisations to create content packages across thousands of journals and provide their customers with content that would have otherwise required an unreasonable amount of time to put together.
How Classify Works
Creating a class, adding concepts (terms) related to the topic of interest, and then perhaps setting up a couple of filters, provides a list of documents that can either be very interesting or in need of refinements by tweaking the list of concepts.
Sounds simple enough, but if you are not a subject expert, how would you add concepts that will get you satisfactory results?
That’s Where Gold Documents Come In
If documents are considered interesting, they can be added as Gold Documents. This will pick up a list of their most essential concepts and base the class results on that list. This way, manually adding specific concepts becomes a step of class refinement instead of the main step.
The BUT About Gold Documents…
As it turns out, one has to be, once again, a subject expert to spot documents that are suitable to be Gold Documents. Fortunately, we have clever users. They found out that if they add a whole issue of a journal, which may consist of 100-150 papers, they will get a very topic-focused list of Gold Documents that can be used.
And there were some problems with that as well: 1. there were too many concepts automatically added which cluttered the classes; and 2. because of the long list of concepts, we had problems processing the results which caused the tool to crash.
What Made It Work
We introduced 3 options. Users could choose between 10, 20, or 40 concepts from Gold Documents to be automatically added as well as the option to tick any concept beyond this threshold if it’s believed to be an important one. We also added an overview of how many times a concept overlaps between all Gold Documents to make it clear if it’s truly an essential one.
Automatically adding the most essential concepts from Gold Documents, while still being able to manually add concepts beyond that threshold enabled users to create content packages easily, as well as it made it easier for us to handle these processes.
A great example of a client who used the tool is OECD. They created their SDG Pathfinder by using Classify.