1
$\begingroup$

Question I was given: Let V be a uniform random variable distributed over the interval (0,1). Let $\ X = \frac{1}{\sqrt(U)}$. What is the cumulative distribution function and probability density function of X?

I understand the basic concept here behind cumulative distribution functions and probability density functions. What's throwing me off here is the additional parameter saying that V is a uniform random variable.

Is the cumulative distance function just: $\ \int\limits_0^1 \frac{1}{\sqrt(V)}dV $ which would then just equal 2?

Then the prob dens function would be: $$ \frac{1}{\sqrt(V)}-----0<V<1 $$ $$ 0------otherwise $$ This seems too simple to me to be the correct answer though. Am I missing something? Am I finding the cumulative distance function of V rather than of X?

$\endgroup$
0
$\begingroup$

We find the cumulative distribution function $F_X(x)$ of $X$. So we want to find $\Pr(X\le x)$.

If $x\le 1$, then $F_X(x)=0$. For since $0\lt U\lt 1$, it is impossible for $\dfrac{1}{\sqrt{U}}$ to be $\le 1$.

Now we find $F_X(x)$ for $x\gt 1$.

We have $X\le x$ if and only if $\dfrac{1}{\sqrt{U}}\le x$. Flip it over. So $X\le x$ if and omly if $\sqrt{U}\ge \dfrac{1}{x}$, that is, if and only if $U\ge \dfrac{1}{x^2}$.

But $\Pr\left(U\ge \dfrac{1}{x^2}\right)=1-\Pr\left(U\le \dfrac{1}{x^2}\right)$. Since $U$ is uniform, this is simply $1-\dfrac{1}{x^2}$. Thus for $X\gt 1$, $$F_X(x)= 1-\frac{1}{x^2}.$$

For the density function $f_X(x)$, differentiate the cdf.

$\endgroup$
  • $\begingroup$ Which would just yield 2/X^3 for X>1 and 0 otherwise? $\endgroup$ – TopherGopher Mar 21 '13 at 2:24
  • $\begingroup$ And thank you - that clarification was fantastic! $\endgroup$ – TopherGopher Mar 21 '13 at 2:24
  • $\begingroup$ You are welcome. Yes, the density function (for $x\gt 1$) is $\frac{2}{x^3}$. $\endgroup$ – André Nicolas Mar 21 '13 at 2:32

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.