MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Hi I'm doing a computation where the modified bessel function of the third kind is the main source of computational strain, we are using a 10,000 dimension's for our distribution, is there any easier way to compute it besides the standard approach, any estimations?

share|cite|improve this question
I think you are using a now obsolete terminology, see Can you tell us what range of parameters you're interested in? – Stopple Oct 10 '11 at 23:20 page 6 – redhatuser Oct 11 '11 at 0:01
What you've termed "the third kind" is what would now be called "the second kind": ; echoing Stopple's question, how big/small are $\nu$ and $z$ in $K_\nu (z)$? How accurate an approximation are you expecting? – J. M. Oct 11 '11 at 8:42
thank you both for clearing that up , z is a 10,000x10,000 matrix, v =5 which is small i was looking for an approximation that would be less computationally complex while still maintaining a high order of accuracy – redhatuser Oct 11 '11 at 20:55
So you have Bessel functions of a matrix argument; the Digital Library of Mathematical Functions has links to papers on asymptotics, see – Stopple Oct 11 '11 at 23:00

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.