From what I've seen in clustering, distance is taken as a hyper parameter (which is to be selected) when inferring the relationships/clusters between points. Examples of distances I've come across are Euclidean, taxicab, Mahalanorbis, and Minoowski.

What are some examples of highly-cited papers doing this, and is there any concrete empirical evidence supporting this practice? And what non-Euclidean distance functions do they use?

  • $\begingroup$ supporting this practice what practice? hyper-parameters tuning? $\endgroup$
    – Alberto
    Commented Aug 17, 2023 at 23:07
  • $\begingroup$ No Turing the distance function as s hyperparameter $\endgroup$
    – ABIM
    Commented Aug 23, 2023 at 3:34


You must log in to answer this question.