Review of Integration and Product Measures

Definitions, Propositions and Theorems

Simple Function

$$\phi (\omega )=\sum _{i=1}^n{a}_i{1}_{A_i}(\omega )$$
where$A_i$ are disjoint sets with finite measure $\mu (A_i)<\infty$.

Proposition 1

Any measurable function $f$ is the limit of a simple function sequence, i.e., $f={\lim }_n{\phi }_n(\omega ).$

Bounded Convergence Theorem

If $X_n\to X$ in prob., $|X_n|<M,\forall n$ where $M$ is a finite constant, then $E[X_n]\to E[X]$.

Fatou’s Lemma

If $X_n\ge 0$, then
$$\lim \underset{n}{\inf }E[X_n]\ge E[\lim \underset{n}{\inf }{X}_n].$$

(入不敷出)

Monotone Convergence Theorem

If $0\le X_n\uparrow X$ then $E[X_n]\to E[X]$.

Dominated Convergence Theorem

If $X_n\to X$ a.s., $X_n<|Y|\forall n$ with $EY<\infty$, then $E[X_n]\to E[X]$.

BCT, Fatou, MCT and DCT are equivalent

In the sense that you can derive the other once you have one.

Pushforward Measure

Let $f$ be a measurable mapping from $({\Omega }_1,{\Sigma }_1,\mu )$ to $({\Omega }_2,{\Sigma }_2)$. Define a measure $\mu \circ f^{-1}$ on $({\Omega }_2,{\Sigma }_2)$:
$$\mu \circ f^{-1}(E)=\mu (f^{-1}(E)),\forall E\in {\Sigma }_2.$$ $\mu \circ f^{-1}$ is called the pushforward measure.

Change of variables formula

Assume that $\mu \circ f^{-1}$ is sigma-finite. Let $\phi$ be any measurable function of $({\Omega }_2,{\Sigma }_2)$, we have
$${\int }_{{\Omega }_1}\phi (f(\omega ))\mu (d\omega )={\int }_{{\Omega }_2}\phi (\omega )\mu \circ f^{-1}(d\omega ).$$

Product Measure

Let $({\Omega }_1,{\Sigma }_1,{\mu }_1)$ and $({\Omega }_2,{\Sigma }_2,{\mu }_2)$ be two sigma-finite measure spaces. Define Ω = Ω 1 × Ω 2 = { ( x , y ) , x ∈ Ω 1 , y ∈ Ω 2 } \Omega=\Omega_1 \times \Omega_2=\{(x,y),x\in\Omega_1,y\in\Omega_2\} Ω=Ω1​×Ω2​={(x,y),x∈Ω1​,y∈Ω2​}, Σ = { A × B , A ∈ Σ 1 , B ∈ Σ 2 \Sigma=\{A\times B, A\in\Sigma_1, B\in\Sigma_2 Σ={A×B,A∈Σ1​,B∈Σ2​}. Let $\mathcal{F}=\sigma (\Sigma )$. Then there is a unique measure $\mu$ on $\mathcal{F}$ with
$$\mu (A\times B)={\mu }_1(A){\mu }_2(B).$$ $\mu$ is called product measure, denoted by ${\mu }_1\times {\mu }_2$.

Fubini’s Theorem

Let $f$ be a measurable function of $(\Omega ,\mathcal{F})$. If $f>0$ or $f$ is integrable, then
$${\int }_{{\Omega }_1}{\int }_{{\Omega }_2}f(x,y){\mu }_2(dy){\mu }_1(dx)={\int }_{\Omega }fd\mu ={\int }_{{\Omega }_2}{\int }_{{\Omega }_1}f(x,y){\mu }_1(dx){\mu }_2(dy).$$

Reference

Durrett, Rick. Probability: theory and examples. Vol. 49. Cambridge university press, 2019.