2
$\begingroup$

I am wondering how I can apply a scale factor to a lognormal distribution with a mean of 1.

Starting with the lognormal distribution $X$~$Lognormal(\mu=-0.5, \sigma=1)$.

Then the mean of the distribution is $e^{\mu+\frac{\sigma^2}{2}}=e^{-0.5+\frac{1^2}{2}}=1$.

This distribution is a normalized version of population distribution. If we know that the population mean is actually 100. How can we scale the lognormal distribution so that it has the same shape, but a mean of 100?

$\endgroup$
1
  • $\begingroup$ Please explain what you might mean by the "same shape." Are you perhaps trying to re-ask this question? $\endgroup$
    – whuber
    Nov 15, 2017 at 22:02

1 Answer 1

3
$\begingroup$

By inspection, $\mu$ is a scale parameter in the lognormal, $\sigma$ is a shape parameter.

Clearly, then, if you want to maintain the same shape and change the scale, you change $\mu$ to $\mu'$ say, such that $e^{\mu'+\frac12\sigma^2}=100$ with the same $\sigma$ as before.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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