This is list of all the methods in unsupervised clustering methods:


and out of these methods, following methods do not take custom pairwise distance matrix.

  • MeanShift
  • KMeans
  • MiniBatchKMeans
  • WARD method (HCA)

I am looking for way(s) to have one of these methods (any of these methods) to take a parameter 'metric' and have it use a 'precomputed' distance matrix.

For example: for DBSCAN has a parameter 'metric' and that parameter takes a 'precomputed' distance matrix.

class sklearn.cluster.DBSCAN(eps=0.5, min_samples=5, metric='precomputed')

I am looking to implement the same as DBSCAN but for any of those above mentioned methods.



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.