I just read a article about the comparison between PCA and hierarchical clustering, but I cannot find the strengths and weakness of clustering compared Principal Component Analysis, what about other clustering algorithm? Could anyone provide some details about this?

PCA - creates a low-dimensional representation of the samples from a data set which is optimal in the sense that it contains as much of the variance in the original data set as is possible. PCA also provides a variable representation that is directly connected to the sample representation, and which allows the user to visually find variables that are characteristic for specific sample groups.

Hierarchical clustering - builds a tree-like structure (a dendrogram) where the leaves are the individual objects (samples or variables) and the algorithm successively pairs together objects showing the highest degree of similarity.

Article: Comparison between clustering and PCA

  • 2
    $\begingroup$ they are solving different problems, what's your question, you want to comparing what? $\endgroup$ – Haitao Du May 7 '18 at 15:37
  • 2
    $\begingroup$ BTW, I personally think they link you provided is not well written. $\endgroup$ – Haitao Du May 7 '18 at 15:42
  • $\begingroup$ Could you explain the relationship between them? $\endgroup$ – FlyingBurger May 7 '18 at 15:44
  • 3
    $\begingroup$ what relationship are you looking for. They do not have clear relationship other than both are unsupervised method. $\endgroup$ – Haitao Du May 7 '18 at 15:45
  • $\begingroup$ These aren't necessarily all that similar. As @hxd1011 notes, they provide answers for different questions. (Also, the linked brochure is terrible; your best bet is to forget anything you read there.) Can you say more about what motivates your question here? Is this just idle curiosity, or are you hoping to apply these to a real dataset to learn something? (What?) Etc. $\endgroup$ – gung - Reinstate Monica May 7 '18 at 18:55

There is a very weak link because both PCA and k-means clustering try to minimize the least squared deviations. But that is a pretty much universal principle, and there exists so much more clustering than just k-means. And does not apply to general hierarchical clustering.

See also: What is the relation between k-means clustering and PCA?

The 'article' you linked, however, should probably go into the trash bin. That is just a sales brochure...

| cite | improve this answer | |
  • 1
    $\begingroup$ +1 "for probably go into the trash bin", you said something I was about to say ... $\endgroup$ – Haitao Du May 7 '18 at 18:24
  • 1
    $\begingroup$ Anony-Mousse, the Q is about hierarchgical clustering, not K-means. I don't think the Q should be closed as "duplicate" therefore (albeit the Q is strange). $\endgroup$ – ttnphns May 7 '18 at 18:39
  • $\begingroup$ My reasoning is that that question shows the link that does exist - k-means is related, not so much hierarchical clustering. But then Ward's HAC is related to k-means... $\endgroup$ – Has QUIT--Anony-Mousse May 7 '18 at 22:28

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.