search

dice-group / LIMES

Link Discovery Framework for Metric Spaces.
https://limes.demos.dice-research.org/
GNU Affero General Public License v3.0
129 stars 54 forks source link

[WebUI] Explorer for Results #232

Open kvndrsslr opened 4 years ago

kvndrsslr commented 4 years ago

In the active learning view, we already have a feature to explore the caches of . It should be possible to reuse this view to explore the results of a linking task.

Todo:

kvndrsslr commented 4 years ago

In Link Discovery, we commonly call the two input datasets Source and Target (S, T). M is the mapping, i.e. the output of LIMES. (s,t)\in M basically means all the pairs of resources s, t that are in LIMES output. Hope that made it clearer. I will have to implement another REST method on the Server. In the meantime, please take the view you already designed for active learning and prepare it for this task by adding pagination. The main difference between this task and the one in active learning is that here the data you get from the server will not be limited to 10 examples. therefore, you need to implement pagination in the client. everything else, including the ability to classify the links as positive, neutral or negative can stay the same for now.