-1
$\begingroup$

Given $X,Y$ real random variables, we know that $E[Y|X]$ is X measurable and that there is a Lebesgue measurable function $f : \mathbf{R} \rightarrow \mathbf{R}$ such that $E[Y|X]=f(X)$ almost everywhere.

Is there any explicit and standard notation for the function $f$ as a function of $X,Y$ ? Something like $\mu_{Y|X}$ (I am just inventing here), so that one could write $\mu_{Y|X}(x_1)$ instead of $f(x_1)$ and therefore make it explicit the meaning of $f$ ?

$\endgroup$
3
  • 1
    $\begingroup$ The standard notation is $E[Y|X]$ and this random variable is defined as a measurable function of $X$ such that$$E[Y]=\int E[Y|x]\,\text{d}P(x)$$ $\endgroup$
    – Xi'an
    Jun 9, 2020 at 11:49
  • $\begingroup$ This looks like the law of total expectation. Why are your reporting it ? Maybe you want to suggest that $E[Y|x]$ is the standard notation for my $f(x)$ ? $\endgroup$
    – Thomas
    Jun 9, 2020 at 11:56
  • $\begingroup$ Not sure why the downvote anyway... $\endgroup$
    – Thomas
    Jun 9, 2020 at 13:13

1 Answer 1

0
$\begingroup$

Same as we can use $f_X(x)$ for the probability density $f$ of random variable $X$ evaluated at $x$, you can use notation like $f_{Y|X}(y|x)$ or $\mu_{Y|X}(y|x)$ to denote conditional things (e.g. densities, expected values etc.). See examples here or here.

$\endgroup$
6
  • $\begingroup$ Great thanks! There is a leap on the notation though, right? Because $Y|X$ is not a well defined random variable. Has $\mu_{Y|X}(y)$ the same meaning of my question ? In this case the natural argument should be $x$ not $y$ or am I wrong ? And how is $f_{Y|X}(y)$ defined ? $\endgroup$
    – Thomas
    Jun 9, 2020 at 10:12
  • $\begingroup$ @Thomas sorry, you're right, updated. $\endgroup$
    – Tim
    Jun 9, 2020 at 10:19
  • $\begingroup$ Thanks Tim, still in my mind it should be $\mu_{Y|X}(x)$ and, if I understand correctly your notation, $f_{Y|X}(y|x)$ is a shorthand for a two-variable function $f_{Y|X}(y,x)$. Than $\mu_{Y|X}(x)$ should have the meaning of the mean of $Y$ along all events where $X=x$, so that depends only on one variable. Whereas (tell me if I am wrong) you are using $y|x$ as a shorthand for $(y,x)$ and $f_{Y|X}(y,x)$ is the probability that $Y=y$ given that $X=x$. We have than $\mu_{Y|X}(x)=\int dy \ y f_{Y|X}(y|x)$ . Is this what you are suggesting ? $\endgroup$
    – Thomas
    Jun 9, 2020 at 12:01
  • $\begingroup$ Sorry but my question is all about notations, what is standard and what is not ... therefore I would like to be precise... $\endgroup$
    – Thomas
    Jun 9, 2020 at 12:04
  • 1
    $\begingroup$ @Thomas there is no strict rules with notations, just conventions as ones mentioned above and by Xi'an in the comment. You can always define what you mean by introducing a notation. $\endgroup$
    – Tim
    Jun 9, 2020 at 12:41

Your Answer

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

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