This problem's been driving me up the walls. I'm not sure where to go with it. Let X be a continuous random variable with density $f_X(x)>0$ for all x. Let $F_X$ be the distribution function for X.

a) Find the density and distribution functions for $U=F_X(X)$.

b) Find the density and distribution functions for $V=log(F_X(X))$.

  • $\begingroup$ $F_V(y) = P(\log X \leq y)=P(X\leq e^y)=F_X(e^y)$ $\endgroup$ – EA304GT Nov 13 '15 at 0:58
  • $\begingroup$ No, @EA304GT $F_V(v) = \mathsf P\Big(\log \big(F_X(X)\big) \leq v\big)$ $\endgroup$ – Graham Kemp Nov 13 '15 at 1:37

We give a brief solution of the first problem.   The second yields to similar tools.   After we know the answer to the first problem, we no longer need to work with the potentially confusing $F_X(X)$.

Let $U=F_X(X)$. Then for $0\lt u\lt 1$, we have $$\begin{align}F_U(u) & =\Pr(U\le u) \\ & =\Pr(F_X(X)\le u) \\ & =\Pr(X\le F_X^{-1}(u)) \\ & =F_X(F_X^{-1}(u)) \\ & = u\end{align}$$

Thus $U$ has uniform distribution on $(0,1)$.

Now find $V= \log U$ ...

  • $\begingroup$ @Graham Kemp: Thanks for the edit, makes it look better, and changing to $U$ is probably a good idea. I have the bad habit of picking my own letters. $\endgroup$ – André Nicolas Nov 13 '15 at 1:35

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.