1
$\begingroup$

It seems they both perform clustering. They both reduce the dimensionality of the input data and classify further inputs based upon their distance/similarity to the center points. These points then update to accommodate the new data.

I am yet to understand how these two methods are different. I suppose it depends on the problem to be solved. How could each be suited to different problems (advantages/disadvantages)?

$\endgroup$
1
  • $\begingroup$ What do they have in common except computing distances? In fact, SOMs are often used as a step before k-means, because SOMs do not produce clusters (instead, you run k-means on the mapped data - SOM itself is not a clustering algorithm!). $\endgroup$ Jun 16, 2018 at 17:56

1 Answer 1

0
$\begingroup$

Self-organizing map (SOM) is a dimensional reduction technique that learns a reduced feature mapping.

Clustering groups individual instances together.

Dimensional reduction and clustering are fundamentally different processes.

$\endgroup$

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.