Use Levenshtein distance to sort LoD Controller results. Much more accurate, less fuzz #171
Labels
enhancement
New feature or request
External API
Things we do not control but we depend on
LoD and Auth Controller
A Tree of Apples
UI/UX
The thing people do when in front of a screen
Milestone
What?
We return results from remote LoD APIs sorted by what the API provides. But really, from the the user/metadata cataloger perspective we want to return things listed by "how close" they are to the original input/search string.
e.g the first term for input "Music" in LoC Subjects is "MUSIC (Computers)"... but the user searched for Music. That is where my little little knowledge of data science comes in place: Levenshtein.
A simple sort-by fixes the issue
See how that/this is better when i use my brainz to add a line of code.
This also helps a lot with unsupervised LoD reconciliation at an AMI level or when using
ami_lod_reconcile(subject|lower|capitalize,'loc;subjects;thing','en',1)
in a twig template.@alliomeria thoughts?
The text was updated successfully, but these errors were encountered: