2
$\begingroup$

Let $(X_n)_{n\in\mathbb{N}}$ be a sequence of independent real valued random variables on the probability space $(\Omega, F, P)$. I'm trying to show that the collection $\mathcal{C}$ of events $\{\omega: X_i(\omega) \in B_i, i = 1,\dots,k\}$, for $B_i \in \mathcal{B}(\mathbb{R})$, generates the sigma-algebra $H_k := \sigma(X_1,\dots,X_k)$. I've already shown that $\mathcal{C}$ is a $\pi$-system hence the next step of my proof would be to show that any $X_i, i = 1,\dots,k$, is $\sigma(\mathcal{C})/\mathcal{B}(\mathbb{R})$ measurable. But given that any element $A \in \mathcal{C}$ has the form $A = \{\omega: X_i(\omega) \in B_i, i = 1,\dots,k\} = \bigcap_{i=1}^k\{\omega: X_i(\omega) \in B_i\} = \bigcap_{i=1}^kX_i^{-1}[B_i]$, I don't really know how to proceed. We certainly know that for any fixed $j$, $A = \bigcap_{i=1}^kX_i^{-1}[B_i] \subset X_j^{-1}[B_j]$, but I can't really convince myself of the fact that then $X_j^{-1}[B_j] \in \sigma(\mathcal{C})$. How should I proceed with this proof?

$\endgroup$

1 Answer 1

2
$\begingroup$

If $B_i=\mathbb R$ for all $i \neq j$ and $B_j=B$ then $\{\omega: X_i(\omega) \in B_i,i=1,2...,k\}=X_j^{-1}(B)$ so $X_j^{-1}(B)\in \mathcal C \subseteq \sigma (\mathcal C)$.

$\endgroup$
2
  • $\begingroup$ While this answer did answer my exact question, could you also give your thought for the converse inclusion. Namely, does it follow from the fact that a similar treatment for other indices shows that $\mathcal{C}$ contains all possible pairs of intersections of all preimages of $X_1,\dots,X_k$, so that $\mathcal{C} \subset H_k \implies \sigma(\mathcal{C}) \subset \sigma(H_k) = H_k$? $\endgroup$ Feb 20, 2022 at 9:54
  • $\begingroup$ Any set in $\mathcal C$ is a finite intersection of sets of the form $X_j^{-1}(B)$ so it is in $H_K$. $\endgroup$ Feb 20, 2022 at 9:57

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .