$X$ is a random variable uniformly distributed on the real interval [0,1].

Through some experimentation, I found that the probability density function, PDF of:

$X$ is $1$ or $\dfrac{d}{dx}X$

$2X$ is $\frac{1}{2}$ or $\dfrac{d}{dx}X/2$

$3X$ is $\frac{1}{3}$ or $\dfrac{d}{dx}X/3$

$X^2$ is $\frac{1}{2\sqrt{X}}$ or $\dfrac{d}{dx}\sqrt{x}$

$X^3$ is $\frac{1}{3x^{2/3}}$ or $\dfrac{d}{dx}\sqrt[3]{x}$

The PDF is useful in answering questions such as what is the mean of $X^3$ or what is the probability that $0<2x<\frac{1}{21}$?

1) How do I find the PDF of functions in general, something like $X+X^3$?

2) Also, when there is another variable involved, say Y that is a random variable uniformly distributed on the real interval [0,2], how do I find the PDF of expressions like $X+Y^2$ or $XY^2$? This is again most helpful in finding answers like what is the variance of $X+Y^2$ or what is the probability that $XY^2 > 1$?

3) What if X and Y are not uniformly distributed, but follows some continuous distribution like the Poisson or Gaussian? How do I find the PDFs in this case?

  • $\begingroup$ There numerous useful posts devoted to this problem: 30938, 77873, 55607 and many others. $\endgroup$ – Sasha Mar 25 '12 at 13:43
  • $\begingroup$ To convince you that this answer is for you, let me quote its first words: The simplest and surest way to compute the distribution density or probability of a random variable is often... $\endgroup$ – Did Mar 25 '12 at 14:21
  • $\begingroup$ possible duplicate of Given the pdf of independent RVs $I$ and $R$, how to find cdf of $W =I^2R$? $\endgroup$ – Did Mar 25 '12 at 14:23
  • $\begingroup$ I apologize but I'm not strong in calculus so I'm still having trouble understanding how to apply your solution in that particular case to the general scenario. Until I do, I'll just leave the question open, hoping someone else can answer it in a different way perhaps. $\endgroup$ – hollow7 Mar 26 '12 at 23:52

You have just discovered that the cumulutative distribution function of an $f(X)$ when $f$ is an invertible monotonuous increasing function can be computed as:

$$\mathbb{P}(f(X)<y)=\mathbb{P}(X<f^{-1}(y)) \; .$$

  • 1
    $\begingroup$ It is conventional to use capital $F$ for the cumulative distribution function and lower-case $f$ for the density function. And I don't think it's a good idea to use the same letter, capital $X$, to refer both to the random variable and to the argument to the CDF. $\endgroup$ – Michael Hardy Mar 25 '12 at 14:57
  • $\begingroup$ $F_Y(y)=\mathbb{P}(Y<y)$ $\endgroup$ – Raskolnikov Mar 25 '12 at 14:59

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.