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

Assume that I have an expectation maximization (EM) trained Gaussian Mixture Model (GMM). So for eg. with three sources i end up with the parameters

 ([u1, u2, u3], [sigma1, sigma2, sigma3], [tau1, tau2, tau3]). 

Is there a way to use this information plus a count of how many data points "d" that were used to make this model to regenerate data that would likely be similar to the original model?

Is it just feeding this Gaussian with random numbers? or is there a better way to do it?

share|cite|improve this question
What's wrong with picking one of the Gaussians w/prob proportional to their weight and sampling from it (all repeated $d$ times)? Why do you need a better way? – Yoav Kallus May 25 '12 at 3:08
I may not, I'm not really sure how to do this which is why I asked, you should add that as an answer. – user23936 May 25 '12 at 5:02

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.