3
$\begingroup$

Is it proper to say that clustering methods are mostly unsupervised learning techniques, with some exceptions such as model-based clustering?

$\endgroup$
  • $\begingroup$ Thanks for your replies. On the other hand, can we consider model-based clustering as supervised learning? (In this context, the data would be "labelled" by introducing the likelihood of belonging to a certain multivariate distribution, which corresponds to a cluster.) $\endgroup$ – Pippo Mar 27 '16 at 15:14
3
$\begingroup$

Clustering is sub-class of unsupervised learning. Unsupervised learning techniques include clustering, feature extraction (e.g., PCA, Isomap, KODAMA), and feature selection (e.g., selection the variables with highest variance value). I think model-based clustering methods are still unsupervised techniques. Although you decide the model what you want, you do not use the information about the labels.

| cite | improve this answer | |
$\endgroup$
1
$\begingroup$

By definition, clustering is unsupervised, with the exception if semi-supervised clustering, where parts of the data are labeled.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

Dimensionality reduction is typically done unsupervised. However, unfortunately, I have seen the term unsupervised learning used synonymously with clustering way too frequently.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

Clustering is generally unsupervised, but there are many other unsupervised learning techniques that are not clustering, per se, for instance the learning of self-organizing feature maps.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.