3
$\begingroup$

Suppose we are given $\mu$ and $\sigma$ for a lognormal distribution with random variable $X$. $\mu$ is the mean of the variable's logarithm and $\sigma$ is the standard deviation of the variable's logarithm. Then if $Z$ is sampled from a standard normal distribution ($Z$ has mean 0 and std deviation 1),

$$X = e^{\mu + \sigma Z}$$

Now, I am given access to a uniform random generator, and use Box-Muller to generate a sample $Z$, can I generate a sample from $X$ by applying the above formula, or will there be issues with doing this?

$\endgroup$
  • $\begingroup$ Yes, that should work fine $\endgroup$ – Glen_b Aug 6 '14 at 23:52
  • $\begingroup$ Yes. You're generating GBM? Correct. $\endgroup$ – SmallChess Aug 7 '14 at 3:20
6
$\begingroup$

X is log-normal means log X is normal. Since the exponential is a monotone-transformation if you correctly sample Z, you will correctly sample X so yes, the method you suggested works

$\endgroup$

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.