Closed georgerichardson closed 1 year ago
I'm happy taking care of this one and exploring the alternatives above.
@ampudia19 - maybe we can do this in a similar method that we discussed for issues 9/10? i.e. we architect it together so the functions take the same inputs/outputs and then we divide and conquer the methods and evaluate them together?
Happily :)
I've implemented the entity count as well as a weighted count of journal relevance per sub-topic (this provides some exogenous labelling of the topics). I'm happy to explore the remaining ones here.
Generate names for DBpedia clusters. At the highest level, we can use manual labelling. At lower levels, we will probably want to use automated methods. Some options include: