FIND ME ON

GitHub

LinkedIn

Filtration

🌱

Definition
MeasureTheoryStochasticProcesses

A filtration on (Ω,F,P)(\Omega,\mathcal{F},P) induced by N\mathbb{N} is a family (Fn)nN(\mathcal{F}_{n})_{n\in\mathbb{N}} of sub σ-algebras of F\mathcal{F} (i.e. Fn\mathcal{F}_{n} is a σ-algebra and FnF, nN\mathcal{F}_{n}\subset \mathcal{F}, \ \forall n\in\mathbb{N}) s.t. FnFn+1, nN\mathcal{F}_{n}\subset \mathcal{F}_{n+1}, \ \forall n\in\mathbb{N}

Let (Xn)nN(X_{n})_{n\in\mathbb{N}} be a Stochastic Process on (Ω,F,P)(\Omega,\mathcal{F},P), nN\forall n\in\mathbb{N}, let the natural Filtration be defined as FnX=σ(X0,,Xn)\mathcal{F_{n}}^{X}=\sigma(X_{0},\dots,X_{n})

The natural filtration is the minimal filtration for any RV XX i.e. (Xn)nN is Fn-adaptive    FnXFn nN(X_{n})_{n\in\mathbb{N}}\text{ is }\mathcal{F}_{n}\text{-adaptive}\implies\mathcal{F}_{n}^{X}\subset \mathcal{F}_{n} \ \forall n\in\mathbb{N}

A filtration (Ft)t0(\mathcal{F}_{t})_{t\ge 0} satisfies the usual conditions if 1. Ft=Ft+\mathcal{F}_{t}=\mathcal{F}_{t^{+}}, t0\forall t\ge 0 where Ft+=s>tFs\mathcal{F_{t^{+}}}=\bigcap_{s>t}\mathcal{F}_{s}i.e. (Ft)t0(\mathcal{F}_{t})_{t\ge 0} is right-continuous and; 2. Ft\mathcal{F}_{t} is complete i.e. AF,BP(Ω),BA  &  P(A)=0    BF (hence P(B)=0)\forall A\in\mathcal{F},\forall B\in\mathcal{P}(\Omega),B\subset A \ \ \& \ \ P(A)=0\implies B\in\mathcal{F} \ (\text{hence }P(B)=0)and; 3. F0\mathcal{F}_{0} contains all elements in F\mathcal{F} of PP-measure 00 i.e. AF s.t. P(A)=0 then AF0\forall A\in\mathcal{F}\text{ s.t. }P(A)=0\text{ then }A\in\mathcal{F}_{0}

Linked from