1
$\begingroup$

I am experimenting with OPTICS clustering in R and from what I have seen in the vignette the valleys and peaks somehow determine the number of clusters which than can be extracted using extractDBSCAN and extractXi. I tried to do clustering on a dataset with 10 features (with minPts = 11) and got the following plot. Could someone help me with it's interpretation, as I do not see any pattern similar to the one in the vignette. enter image description here

Improve this question
$\endgroup$

1 Answer 1

Reset to default
1
$\begingroup$

A shape like this indicates that you have some density peaks (these are easy to spot...). But you don't appear to have larger areas of similar high density, which would cause wider valleys.

Whether or not these narrow density peaks are to be considered clusters depends on your data and task.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Could you give an example when it would be sufficient to form cluster on this basis? I am doing some customer segmentation with multiple features per client. $\endgroup$
    – abu
    Commented Apr 4, 2018 at 20:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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