1
  • In Fourier series, any function can be decomposed as sum of sine and cosine
  • In neural networks, any function can be decomposed as weighted sum over logistic functions. (A one layer neural network)
  • In wavelet transforms, any function can be decomposed as weighted sum of Haar functions

Is there also such property for decomposition into mixture of Gaussians? If so, is there a proof?

Rob Romijnders
  • 833
  • 1
  • 6
  • 16
  • Wavelet transforms are much more than just Haar functions. – DaBler May 09 '19 at 12:37
  • 2
    You may also want to check out https://math.stackexchange.com/ and https://mathoverflow.net/. I tried to edit my answer to give more context, but it is hard to do here because there's no latex/mathjax in the markdown. If you're interested, you can ask a similar question there and I'll try to answer? – Riley Jul 06 '20 at 19:26

3 Answers3

4

There's a theorem, the Stone-Weierstrass theorem, which gives conditions for when a family of functions can approximate any continuous function. You need

  • an algebra of functions (closed under addition, subtraction, and multiplication)

  • the constant functions

  • and you need the functions to separate points:

    • (for any two distinct points you can find a a function that assigns them different values)

You can approximate a constant function with increasingly wide gaussians. You can time-shift gaussians to separate points. So if you form an algebra out of gaussians, you can approximate any continuous function with them.

Riley
  • 982
  • 1
  • 7
  • 19
  • +1 This answer is significantly more thorough than the top-voted answer. I've referenced your answer [here](https://mattermodeling.stackexchange.com/questions/9681/do-the-cc-pc-def2-basis-sets-mathematically-converge-to-the-cbs-limit-assuming/9682?noredirect=1#comment20415_9682) in case you're interested! Perhaps you can write a better answer for that user who seems to be much more interested in the mathematical convergence properties than the typical user of that Stack Exchange site? – Nike Sep 23 '22 at 18:56
3

If the sum allows to be infinite, then the answer is Yes. Please refer to Yves Meyer's book of "Wavelet and Operators", section 6.6, lemma 10.

  • Note that this lemma is for complex functions expressed as sum of positive Gaussians with complex coefficients, so it might take a bit more work to show (if true) that any positive function can be written as a sum of positive Gaussians with positive coeffcients. – Jess Riedel Sep 13 '20 at 18:00
1

Yes. Decomposing any function to a sum of any kind of Gaussians is possible, since it can be decomposed to a sum of Dirac functions :) (and Dirac is a Gaussian where the variance approaches zero).

Some more interesting questions would be:

  • Can any function be decomposed to a sum of non-zero variance Gaussians, with a given, constant variance, that are defined around varying centers?
  • Can any function be be decomposed to a sum of non-zero variance Gaussians, all having 0 as the center, but defined with alternating variances?

The Mathematics Stack Exchange might be a better place to answer these questions though.

Nike
  • 1,223
  • 2
  • 19
  • 42
ntg
  • 12,950
  • 7
  • 74
  • 95