To calculate the expectation of $e^x$ for a standard normal distribution I eventually get, via exponential simplification:


When I plug this into Wolfram Alpha I get $\sqrt e$ as the result. I'd like to know the integration step(s) or other means I could use to obtain this result on my own from the point I stopped. I am assuming that Wolfram Alpha "knew" an analytical solution since it presents $\sqrt e$ as a solution as well as the numerical value of $\sqrt e $. Thanks in advance!

  • $\begingroup$ WolframAlpha only shows steps for indefinite integrals; you can often use these to derive a result on a definite integral. This particular function has no elementary anti-derivative, however. You have $\sqrt{e}$ times a normal distribution and the integral of the normal is well known to be 1. $\endgroup$ – Mark McClure May 31 '13 at 0:00
  • $\begingroup$ Thanks for responding Mark. That actually helps. Since it started as $e^x$ multiplied by the normal probability density function, which must equal one, I should have known that once I got the $\sqrt{e}$ to pop out through simplification, all that was left was the pdf. Appreciate you helping me see that. $\endgroup$ – Joe May 31 '13 at 0:29

This is because $$\int_{\Bbb R} e^{-x^2}dx=\sqrt \pi$$

Note that your shift $x\mapsto x-1$ doesn't change the value of integral, while $x\mapsto \frac{x}{\sqrt 2}$ multiplies it by $\sqrt 2$, giving the desired result, that is,

$$\int_{\Bbb R} e^{-\frac 1 2(x-1)^2}dx=\sqrt {2\pi}$$

  • $\begingroup$ Thanks Peter! I'd give you points, but I am new here. $\endgroup$ – Joe May 31 '13 at 0:00
  • $\begingroup$ @Joe When you get some upvotes for your answer, you will be able to upvote, no problem. $\endgroup$ – Pedro Tamaroff May 31 '13 at 0:02

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.