Machine Learning

1 readers

1 users here now

Community Rules:

Be nice. No offensive behavior, insults or attacks: we encourage a diverse community in which members feel safe and have a voice.
Make your post clear and comprehensive: posts that lack insight or effort will be removed. (ex: questions which are easily googled)
Beginner or career related questions go elsewhere. This community is focused in discussion of research and new projects that advance the state-of-the-art.
Limit self-promotion. Comments and posts should be first and foremost about topics of interest to ML observers and practitioners. Limited self-promotion is tolerated, but the sub is not here as merely a source for free advertisement. Such posts will be removed at the discretion of the mods.

founded 1 year ago

MODERATORS

communick@academy.garden

[D] Unsupervised Clustering without knowing number of classes (alien.top)

submitted 1 year ago by BigBrainUrinal@alien.top to c/machinelearning@academy.garden

5 comments fedilink hide all child comments

Does anyone know where to find the best models for unsupervised clustering problems that don't specify the number classes? For example I googled unsupervised MNIST but IIC which holds the record requires the output dimension (k=10) to be specified? Is there a name for unsupervised clustering without knowing the number of classes? (I know of density/hierarchical clustering algorithms but am unaware of many deep learning ones) And specifically are results charted anywhere? I'm researching the topic and it seems knowing the number of things you're looking for is half the battle. I can find papers on methods that aim to find the number of clusters etc but are there any benchmarks to compare?

you are viewing a single comment's thread
view the rest of the comments

[–] visarga@alien.top 1 points 1 year ago

You already mentioned hierarchical methods but I got my best class-count agnostic clustering with fclusterdata from scipy:

labels = fclusterdata(data, t=threshold, criterion='distance')