2
$\begingroup$

I would like to plot the PDF of the product of a $Beta(1,1)$ and $Beta(3/2,1/2)$ random variables. I can use the following to calculate the PDF at a point $x$:

PDF[TransformedDistribution[u v, {u \[Distributed] BetaDistribution[1, 1], 
v \[Distributed] BetaDistribution[3/2, 1/2]}], x]

However, to work out the PDF at a point (take $x=0.5$ for example), the above takes about 10 minutes on my computer.

I could leave my computer overnight, and it should be able to plot the above PDF using $Plot$, however, I was wondering whether there are any quicker alternatives?

Note: I want to plot the distribution's PDF exactly, not an approximation by drawing randomly first a $Beta(1,1)$ then multiplying it by a draw from $Beta(3/2,1/2)$, then looking at a histogram.

Best,

Ben

Improve this question
$\endgroup$

1 Answer 1

Reset to default
5
$\begingroup$

From what I can tell you are recalculating the transformed distribution too often in your plot. Calculate and store the resulting PDF once and then use it for your plots.

tpdf = PDF[
  TransformedDistribution[
   u v, {u \[Distributed] BetaDistribution[1, 1], 
    v \[Distributed] BetaDistribution[3/2, 1/2]}], x]

Once you have tpdf your plot will return immediately since you are not recalculating it for each value of x.

Plot[tpdf, {x, 0, 1}]

enter image description here

Hope this helps.

Improve this answer
$\endgroup$
4
  • 2
    $\begingroup$ …and if OP would rather skip the use of an auxiliary variable, there's the option setting Evaluated -> True. $\endgroup$
    – J. M.'s lack of A.I. ♦
    Aug 11, 2015 at 16:41
  • $\begingroup$ @Guesswhoitis. Is this an option on Plot? I can't seem to find it in 10.2 $\endgroup$
    – Edmund
    Aug 11, 2015 at 17:21
  • $\begingroup$ @Edmund - use Plot[Evaluate[PDF[..., x]], {x, 0, 1}] $\endgroup$
    – Bob Hanlon
    Aug 11, 2015 at 18:49
  • $\begingroup$ Yes, it's an option for Plot[]. $\endgroup$
    – J. M.'s lack of A.I. ♦
    Aug 12, 2015 at 0:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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