0
$\begingroup$

I am looking for a concise description to help me understand the error for Monte Carlo Integration using Uniform Sampling and Importance Sampling

For Importance Sampling I have that the error is just the square root of the variance...

$$\sqrt{\frac{1}{N}<f(Xi)^2>-<\frac{1}{N}f(Xi)>^2)}$$

for Uniform Sampling what would the error be similar such as

$$\frac{b-a}{\sqrt{N}}\sqrt{\frac{1}{N}<f(Xi)^2>-<\frac{1}{N}f(Xi)>^2)}$$

where $a$ and $b$ are the limits of the integration??

This to me isn't right as my understanding was that an importance sampling would decrease the variance of the MC integration.

I can't find any clear explanation of the difference online so any help is appreciated.

| cite | improve this question | |
$\endgroup$
  • 1
    $\begingroup$ I can't really answer your question, but it seems odd to me that there are no integration bounds in your first formula. Don't you need them in any case, no matter the sampling strategy? $\endgroup$ – MPIchael Oct 23 at 7:54
  • $\begingroup$ In both of the formulas you show, there is a factor $1/N$ that appears once within the squares and once outside the squares. This seems wrong to me. $\endgroup$ – Wolfgang Bangerth Oct 23 at 14:15
  • $\begingroup$ Your question would be clearer if you cited the sources of these formulas. $\endgroup$ – Brian Borchers Oct 27 at 4:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.