0
$\begingroup$

Having some trouble conceptualizing this. Suppose we have an underlying set $\Omega$, a sigma algebra of subsets $F$, two random variables X and Y on $\Omega$, and two probability measures $P_1$ and $P_2$ where X and Y are independent with respect to $P_1$ but not $P_2$. What is an example of this? I'm honestly having trouble understanding the idea of being independent on one probability measure but not the other.

$\endgroup$
1
$\begingroup$

Let $X$ and $Y$ be independent random variables with, say Poisson (1) distribution, $A=\{X=Y\}$ and $Q(E)=\frac {P(A\cap E)} {P(A)}$. Then $X=Y$ almost surely w.r.t. $Q$ so they are not independent w.r.t. $Q$. [It should be noted that $X$ and $Y$ are not constants w.r.t. $Q$. In fact they take all non-negative integer values with positive probability].

$\endgroup$
0
$\begingroup$

Suppose $X$ and $Y$ are discrete random variables taking values in finite spaces $\mathcal{X}$ and $\mathcal{Y}$ respectively. Then you can consider $\Omega = \mathcal{X} \times \mathcal{Y}$ with the discrete $\sigma$-algebra. Specifying a probability measure on $\Omega$ is essentially filling in a table of probabilities, i.e. specifying a probability for each pair $(x,y) \in \Omega$, and you write down explicitly what independence looks like. You can pick probabilities to satisfy the independence assumption, or you can pick them to violate the independence assumption.

$\endgroup$
  • $\begingroup$ Thanks! I picked $\Omega=\{1,2\}$ and just defined all my individual and joint probabilities arbitrarily. $\endgroup$ – Jax Sep 11 '18 at 8:03

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.