Mehmet Emre Tiryaki
Back to Notes

Sets and σ-fields

Sets

P(Ω)=2Ω={A:A⊆Ω}\mathcal{P}(\Omega) = 2^\Omega = \{A : A \subseteq \Omega\} AB=A∩B,A△B=(A∖B)∪(B∖A)AB = A \cap B, \qquad A \triangle B = (A \setminus B) \cup (B \setminus A) 1A(ω)={1,ω∈A0,ω∉A\mathbf{1}_A(\omega) = \begin{cases} 1, & \omega \in A \\ 0, & \omega \notin A \end{cases}

Limits of Sets

The inner and outer tails of a sequence A1,A2,…A_1, A_2, \ldots are

inf⁡k≥nAk=⋂k=n∞Ak,sup⁡k≥nAk=⋃k=n∞Ak\inf_{k \ge n} A_k = \bigcap_{k=n}^{\infty} A_k, \qquad \sup_{k \ge n} A_k = \bigcup_{k=n}^{\infty} A_k lim inf⁡n→∞An=⋃n=1∞⋂k=n∞Ak=lim⁡n→∞inf⁡k≥nAk\liminf_{n \to \infty} A_n = \bigcup_{n=1}^{\infty} \bigcap_{k=n}^{\infty} A_k = \lim_{n \to \infty} \inf_{k \ge n} A_k

This is the set of ω\omega in some inner tail, so {n:ω∉An}\{n : \omega \notin A_n\} is finite. After some nn, ω\omega is in all AkA_k, i.e. ω\omega belongs to all but finitely many AkA_k.

lim sup⁡n→∞An=⋂n=1∞⋃k=n∞Ak=lim⁡n→∞sup⁡k≥nAk\limsup_{n \to \infty} A_n = \bigcap_{n=1}^{\infty} \bigcup_{k=n}^{\infty} A_k = \lim_{n \to \infty} \sup_{k \ge n} A_k

This is the set of ω\omega in all outer tails, so {n:ω∈An}\{n : \omega \in A_n\} is infinite. It is the set of outcomes in infinitely many AnA_n, written {An i.o.}\{A_n \text{ i.o.}\}. Equivalently,

lim sup⁡n→∞An={ω:∑j=1∞1Aj(ω)=∞}\limsup_{n \to \infty} A_n = \Big\{\omega : \sum_{j=1}^{\infty} \mathbf{1}_{A_j}(\omega) = \infty\Big\}

Always lim inf⁡An⊆lim sup⁡An\liminf A_n \subseteq \limsup A_n, and lim⁡n→∞An\lim_{n \to \infty} A_n exists if both agree.

For monotone sequences the limit always exists. If A1⊆A2⊆⋯A_1 \subseteq A_2 \subseteq \cdots then lim⁡An=⋃nAn\lim A_n = \bigcup_{n} A_n, written An↑AA_n \uparrow A. If A1⊇A2⊇⋯A_1 \supseteq A_2 \supseteq \cdots then lim⁡An=⋂nAn\lim A_n = \bigcap_{n} A_n, written An↓AA_n \downarrow A.

Classes of Sets

  • Field Nonempty, closed under complements, finite unions and finite intersections.
  • σ\sigma-field Nonempty, closed under complements, countable unions and countable intersections. Equivalently ∅∈C\emptyset \in \mathcal{C}, A∈C⇒Ac∈CA \in \mathcal{C} \Rightarrow A^c \in \mathcal{C}, and A1,A2,…∈C⇒⋃n=1∞An∈CA_1, A_2, \ldots \in \mathcal{C} \Rightarrow \bigcup_{n=1}^{\infty} A_n \in \mathcal{C}.
  • π\pi-system Nonempty, closed under finite intersections.
  • λ\lambda-system Contains Ω\Omega, closed under complements and countable disjoint unions.
  • Semialgebra A π\pi-system containing ∅,Ω\emptyset, \Omega such that for each A∈SA \in \mathcal{S} there are pairwise disjoint C1,…,Cn∈SC_1, \ldots, C_n \in \mathcal{S} with Ac=⋃k=1nCkA^c = \bigcup_{k=1}^{n} C_k.

Generated σ\sigma-field

An intersection of σ\sigma-fields is a σ\sigma-field. For C⊆P(Ω)\mathcal{C} \subseteq \mathcal{P}(\Omega), let

H={B:B is a σ-field on Ω, C⊆B},σ(C)=⋂B∈HB\mathcal{H} = \{\mathcal{B} : \mathcal{B} \text{ is a } \sigma\text{-field on } \Omega,\ \mathcal{C} \subseteq \mathcal{B}\}, \qquad \sigma(\mathcal{C}) = \bigcap_{\mathcal{B} \in \mathcal{H}} \mathcal{B}

σ(C)\sigma(\mathcal{C}) is the smallest σ\sigma-field containing every set in C\mathcal{C}.

The Borel σ\sigma-field on R\mathbb{R} is

B(R)=σ({(a,b):−∞<a<b<∞})\mathcal{B}(\mathbb{R}) = \sigma(\{(a,b) : -\infty < a < b < \infty\})