5
$\begingroup$

Suppose one has a distribution of $N$ points on the sphere. Is there an agreed upon metric for the difference of this distribution and $N$ equidistant points on the sphere? To me entropy seems like the right concept, but I'm not sure how to go about defining it.

$\endgroup$

1 Answer 1

3
$\begingroup$

there exists a great variety of metrics for the uniformity of points on a (hyper)sphere; the reference list of this 2010 paper will point you to them. For a comparison of the different metrics:

  • A. Figueiredo, Comparison of tests of uniformity defined on the hypersphere, Statistics and Probability Letters, 77, 329 (2007).
  • P.J. Diggle, N.I. Fisher, and A.J. Lee, A comparison of tests of uniformity for spherical data, Australian Journal of Statistics, 27, 53 (1985).
$\endgroup$
1
  • 1
    $\begingroup$ In other words, there is no unique concept of "N equidistant points on the sphere," except for a few very symmetric cases where the metrics agree. $\endgroup$
    – user25199
    Sep 5, 2013 at 11:35

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.