I am slightly confused about the use of indicator variables in probability density functions.

For example, consider the density of $(X,Y)$ uniformly distributed on the unit disc. This density can be written as $$f_{X,Y}(x,y)= \frac{1}{\pi} \mathbb{1} \left \{ (x,y) \mid x^2 + y^2 \leq 1 \right \}.$$

However, I am not really seeing how this makes sense. I understand indicator variables when solving problems with probability, but why not just write the density without it? What is it exactly saying?

Thanks in advance.


Note that $f_{X,Y}:\mathbb R^2\to\mathbb R$.

It says exactly that $f_{X,Y}(x,y)$ will take value $\frac1{\pi}$ if $x^2+y^2\leq1$ and will take value $0$ otherwise.

So an excellent presentation of a PDF which is formally for a fixed nonnegative integer $n$ a nonnegative measurable function $\mathbb R^n\to\mathbb R$ that gives value $1$ by integration wrt Lebesgue-measure.

In notation it can be handsome to write things like: $\mathbb EX=\int f_X(x)xdx$ without bothering on borders.

If it is not your taste then of course you can also choose for discerning cases: $x^2+y^2\leq1$ and "otherwise" without any mentioning of indicatorfunction.


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.