We have a large amount (Billions) of high cardinality, mixed nominal & numerical data, and are performing some clustering on it as an experiment. There is a small subset of these data, however, that we know should belong in its own cluster.

Is it possible to purposely make sure these data end up in their own cluster, while still clustering the remaining data? If so, how does one approach doing so?

  • $\begingroup$ Why not remove the data in question and run a clustering algorithm on the rest of the data? $\endgroup$ Apr 6, 2018 at 2:26
  • 1
    $\begingroup$ @timleathart well that doesn't really guarantee that the rest of the data will be pushed to a cluster of their own. This is a tough question since one could potentially introduce more than one bias. $\endgroup$ Apr 6, 2018 at 7:01

1 Answer 1


In constrained clustering you can provide examples of objects that should, or that must not, be in the same cluster.

This can be used, e.g., for model selection: run several times and return the result with the fewest violations. Or to guide cluster extraction from a hierarchical result.


Your Answer

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

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