A question about applying strong law of large numbers

86 Views Asked by At

Let $p, q$ be two strictly-positive probability mass functions on a finite space $\Omega$. Let $X_1, X_2 \cdots, \overset{i.d.d}{\sim} q$ and $Y \sim p $. Let $f: \Omega \to \mathbb R$. Define $\omega(x) = \frac{p(x)}{q(x)}$. Show that $\frac{\sum\limits_{i=1}^nf(X_i)\omega(X_i)}{\sum\limits_{i=1}^n\omega(X_i)} \to E[f(Y)]$ almost surely.

For the definition of probability mass function, see


This is a Comprehensive exam practice problem given by our instructor but to be honest, it looks quite terrifying to me, what's the idea of how to approach and solve this problem?

1

There are 1 best solutions below

5
On BEST ANSWER

Expanding on my hint a little bit: after you divide top and bottom by $n$, the top is the sample mean of $f(X) \omega(X)$ and the bottom is the sample mean of just $\omega(X)$. By SLLN both those sample means converge. What do they converge to?