0
$\begingroup$

Let $(\Omega,\mathscr{F}, P)$ be a probability space, let $\nu$ be a finite measure on $\mathscr{F}$, and let $\mathscr{F}_{1}$, $\mathscr{F}_{2}$,... be a non-decreasing sequence of $\sigma$-fields in $\mathscr{F}$. Suppose that $P$ dominates $\nu$ when both are restricted to $\mathscr{F}_{n}$. There is then a density or Radon-Nikodym derivative $X_{n}$ of $\nu$ with respect to $P$ (the $X_{n}$ are a martingale with respect to the $\mathscr{F}_{n}$).

Now, let $P$ be Lebesgue measure on the $\sigma$-field $\mathscr{F}$ of Borel subsets of $\Omega = (0, 1]$, and let $\mathscr{F}_{n}$ be the finite $\sigma$-field generated by the partition of $\Omega$ into dyadic intervals $(k2^{-n},(k+1)2^{-n}],0\leq k \leq 2^{n}$. we know that for every finite measure $\nu$ the Radon-Nikodym derivative is $$X_{n}(\omega)=\dfrac{\nu(k2^{-n},(k+1)2^{-n}]}{2^{-n}}$$ where $\omega\in(k2^{-n},(k+1)2^{-n}]$. But I don't know how to derive it.

Thanks and Regards!

$\endgroup$
2
$\begingroup$

Hints:

  1. Show (or recall) that any $\mathcal{F}_n$-measurable function $X$ is of the form $$X(\omega) = \sum_{j=0}^{2^n-1} c_j 1_{(j 2^{-n},(j+1)2^{-n}]}(\omega)$$ for suitable constants $c_j \in \mathbb{R}$.
  2. Since $X_n$ is $\mathcal{F}_n$-measurable, Step 1 shows that there exist constants such that $$X_n(\omega) = \sum_{j=0}^{2^n-1} c_{j,n} 1_{(j 2^{-n},(j+1)2^{-n}]}(\omega).$$
  3. If $X_n$ is the Radon-Nikodym derivative of $\nu$ with respect to $P$ restricted to $\mathcal{F}_n$, it holds that $$\int_F X_n \, d\mathbb{P} = \nu(F)$$ for all $F \in \mathcal{F}_n$. Choosing $F := (j2^{-n},(j+1)2^{-n}] \in \mathcal{F}_n$, we find $$c_{j,n} 2^{-n} = \int_F X_n \, d\mathbb{P}= \nu(F) = \nu((j 2^{-n},(j+1)2^{-n}]),$$ i.e. $$c_{j,n} = \frac{\nu((j 2^{-n},(j+1)2^{-n}])}{2^{-n}}.$$ Hence, $$X_n(\omega) = \sum_{j=0}^{2^n-1} \frac{\nu((j 2^{-n},(j+1)2^{-n}])}{2^{-n}}1_{(j 2^{-n},(j+1)2^{-n}]}(\omega).$$
$\endgroup$

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.