0
$\begingroup$

I am stuck with the following problem. If we let $Z\sim \mathcal{N}(\mu,\sigma^2)$ and let $F(z)$ denote its CDF, and denote by $\Phi(x)$ the CDF of the standard normal distribution, how can we express integrals of the type $$\int_{-\infty}^a t(z)dF(z)$$ for some function $t(z)$ in terms of $\Phi(x)$?

For example for $t(z)=1$ we get $$\int_{-\infty}^a dF(z)=P(Z\leq a)=\Phi\left(\frac{a-\mu}{\sigma}\right)$$ but what about the more general case? Is there a way to do this? Specifically I am interested in the case $t(z)=e^z$. I hope the question makes sense.

$\endgroup$
1
$\begingroup$

Define: $$U:=\frac{Z-\mu}{\sigma}$$ Then $U$ has standard normal distribution, and for any suitable function $g$ we have: $$\int g(z)dF(z)=\mathbb Eg(Z)=\mathbb Eg(\mu+\sigma U)=\int g(\mu+\sigma u)d\Phi(u)$$

Observe that $$\int_{-\infty}^at(z)dF(z)=\int g(z)dF(z)$$ for $g(z):=t(z)1_{(-\infty,a]}(z)$

$\endgroup$
1
  • $\begingroup$ of course, thanks! $\endgroup$
    – user128836
    Feb 23 '17 at 16:15

Your Answer

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

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