Choose how many clusters you want to generate (i.e. how many centers). You can
select:
- A large number of centers (in which case you’ll need to decide how
many to actually display; there are limitations on how many words and
clusters you can view in the console in RStudio)
- A small number of centers (in which case you can look at them all)
Choose how many randomly selected clusters to actually view
Choose how many words from your randomly selected clusters to see
Run several times: how do the large vs. small number of centers give you a
different view of your corpus?
Are there any clusters that seem particularly interesting, give you insight
into the composition of the corpus (or maybe its data preparation)?
Are there any words or semantic spaces that you might want to explore in more
detail?
Word Vectors: Hands-on Practice and Group Work, Intensive
Pedagogy-focused slide 03 of 11
© 2019 Syd Bauman, Julia Flanders, Sarah Connell, and the Women Writers Project This
TEI-encoded XML file is available under the terms of the Creative Commons Attribution-ShareAlike
3.0 (Unported) license.