Closed aitrain closed 3 years ago
They serve as a sanity check on if the co-occurrences are aligned with your mental model. For example, in a dataset if forks and spoons co-occur frequently, this is probably what you would expect, but if forks and soccer balls also co-occur just as frequently, this may lead to a different understanding of what kinds of images the dataset contain (maybe more outdoor eating scenes than indoor ones)
Hi Angelina,
just a quick question. How should the hierarchy visualization be interpreted?
Thank you very much in advance! :)
All the best, Nastassia