A Probability Path.pdf

# Corollary 5 71 sections of measurable functions are

• Test Prep
• 464

This preview shows pages 160–164. Sign up to view the full content.

Corollary 5. 7.1 Sections of measurable functions are measurable. That is, if then Proof. Since X is 81 x 82/ S measurable, we have for A e S that and hence by the previous results However (X- 1 (A))w 1 ={w2 : X(wt, W2) E A} ={W2 : Xw 1 (w2) E A}= (Xw 1 )- 1 (A), which says Xw 1 is 82/S measurable. 0

This preview has intentionally blurred sections. Sign up to view the full version.

5.8 Probability Measures on Product Spaces 147 5.8 Probability Measures on Product Spaces We now consider how to construct probability measures on the product space (Qt x !lz, Bt x Bz). In particular, we will see how to construct independent random variables. Transition Functions. Call a function a transition function if (i) for each Wt, K(wt. ·)is a probability measure on Bz, and (ii) for each Az e Bz, K(·, Az) is Bt/8([0, 1]) measurable. Transition functions are used to define discrete time Markov processes where K (cut. Az) represents the conditional probability that, starting from Wt, the next movement of the system results in a state in Az. Here our interest in transi- tion functions comes from the connection with measures on product spaces and Fubini's theorem. Theorem 5.8.1 Let Pt be a probability measure on Bt, and suppose K : !lt x Bz [0, 1] is a transition function. Then K and Pt. uniquely determine a probability on Bt x Bz via the formula P(At x Az) = { K (cut. Az)Pt (dwt), (5.20) }AI fora// At x Az E RECT. The measure P given in (5.20) is specified on the semialgebra RECT and we need to verify that the conditions of the Combo Extension Theorem 2.4.3 on page 48 are applicable so that P can be extended to a(RECT) = Bt x Bz. We verify that Pis a-additive on RECT and apply the Combo Extension The- orem 2.4.3. Let { A(n) x A(n) n > 1} t 2 ' - be disjoint elements of RECT whose union is in RECT. We show 00 00 P(L X = L X n=t n=t Note if Ln x =At x Az, then 1A 1 (wt)1A 2 (wz) = \ 1A 1 xA2(Wt. wz) = L 1 wz) n = L1Atn)(Wt)lA<nJ(lUz). n 1 2
148 5. Integration and Expectation Now making use of the series form of the monotone convergence theorem we have the following string of equalities: P(At x A2) = { 1A 1 (wt)K(wt, A2)P1(dwt) ln1 = { [ { 1A 1 (WI)lA 2 (W2)K(wi,dW2)]Pt(dwt) ln1 ln2 = { [ { L)A<n>(wt)lA<n>(W2)K(wt,dW2)]Pt(dwt) lnl lnz n I 2 = { L) { lA<nJ(WI)lA<n>(W2)K(wt.dW2)]Pt(dwt) n I 2 = L { lA(nJ(Wt)[ { 1A<nJ(W2)K(Wt,dW2)]Pt(dwt) n lnl I ln2 2 = L { n lnl I = L 1 K (WI, An))Pt (dwt) n A\n) = L X Ain)). n 0 Special case. Suppose for some probability measure P2 on 82 that K (wt. A2) = P 2 (A 2) . Then the probability measure P, defined in the previous result on Bt x Bz is P(At x A2) = Pt(At)Pz(Az) . We denote this P by Pt x Pz and call P product measure . Define u-fields in nt x n2 by B'f = {At x nz :At e 8t} = {Qt X A2 : A2 E 82} . With respect to the product measure P = Pt x P2, we have l3lj JL since P(At x n2 n n1 x A2) = P(At x A2) = P1 (At)P2(A2) = P(At x n2)P(nt x A2). Suppose X ; : (Q;, 8;) (IR, 8(1R)) is a random variable on n; fori= 1, 2. Define on S'2t X n2 the new functions

This preview has intentionally blurred sections. Sign up to view the full version.

5.9 Fubini's theorem 149 With respect toP = Pt x P2, the variables Xf and are independent since P[Xf::; x, y] = Pt x P2({(w1, w2): Xt (wt) ::; x, X2(W2) :=: y}) = Pt x P2({w1: Xt(wt)::::: x} x {W2: X2(W2)::::: y}) = Pt({Xt(wt)::::: x})P2({W2: X2(W2)::::: y}) = P({(wt. W2): Xf(wt. W2)::::: x}) P({(wt, W2): W2) ::; y}) = P[Xf ::: X ::: y ].
This is the end of the preview. Sign up to access the rest of the document.
• Fall '08
• Staff

{[ snackBarMessage ]}

### What students are saying

• As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

Kiran Temple University Fox School of Business ‘17, Course Hero Intern

• I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

Dana University of Pennsylvania ‘17, Course Hero Intern

• The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

Jill Tulane University ‘16, Course Hero Intern