Monotone class theorem: Difference between revisions

Revision as of 13:24, 6 August 2014

Latest revision as of 01:03, 19 March 2025

(35 intermediate revisions by 18 users not shown)

Line 1:

{{Short description|Measure theory and probability theorem}}

In [[Measure (mathematics)|measure theory]] and [[Probability theory|probability]], the '''monotone class theorem''' connects monotone classes and [[Sigma-algebra|{{sigma}}-algebra]]s. The theorem says that the smallest [[#Definition of a monotone class|monotone class]] containing an [[fieldField of sets|algebra of sets]] <math>G</math> is precisely the smallest [[Sigma-algebra|σ{{sigma}}-algebra]] containing &nbsp;<math>G.</math> It is used as a type of [[transfinite induction]] to prove many other theorems, such as [[Fubini's theorem]].

==Definition of a monotone class==

A '''monotone class''' in a set <math>R</math> is a collection <math>M</math> of [[subsets]] of <math>R</math> which contains <math>R</math> and is [[Closure (mathematics)|closed]] under countable monotone unions and intersections, i.e. if <math>A_i \in M</math> and <math>A_1 \subset A_2 \subset \ldots</math> then <math>\cup_{i = 1}^\infty A_i \in M</math>, and similarly for intersections of decreasing sequences of sets.

A '''{{em|{{visible anchor|monotone class}}}}''' is a [[Family of sets|family]] (i.e. class) <math>M</math> of sets that is [[Closure (mathematics)|closed]] under countable monotone unions and also under countable monotone intersections. Explicitly, this means <math>M</math> has the following properties:

==Monotone class theorem for sets==

# if <math>A_1, A_2, \ldots \in M</math> and <math>A_1 \subseteq A_2 \subseteq \cdots</math> then <math display="inline">{\textstyle\bigcup\limits_{i = 1}^\infty} A_i \in M,</math> and

===Statement===

# if <math>B_1, B_2, \ldots \in M</math> and <math>B_1 \supseteq B_2 \supseteq \cdots</math> then <math display="inline">{\textstyle\bigcap\limits_{i = 1}^\infty} B_i \in M.</math>

Let G be an [[field of sets|algebra of sets]] and define M(G) to be the smallest monotone class containing G. Then M(G) is precisely the [[Sigma-algebra|σ-algebra]] generated by G, i.e. σ(G) = M(G)

==Monotone class theorem for sets==

===Proof===

The following was taken from Probability Essentials, by Jean Jacod and Philip Protter.<ref name="Jacod">{{cite book|last=Jacod|first=Jean|author2=Protter, Phillip |year=2004|title=Probability Essentials|publisher=Springer|page=36|isbn=978-3-540-438717}}</ref> The idea is as follows: we know that the sigma-algebra generated by an algebra of sets G contains the smallest monotone class generated by G. So, we seek to show that the monotone class generated by G is in fact a sigma-algebra, which would then show the two are equal.

{{math theorem|name=Monotone class theorem for sets|note=|style=|math_statement=

To do this, we first construct monotone classes that correspond to elements of G, and show that each equals the M(G), the monotone class generated by G. Using this, we show that the monotone classes corresponding to the other elements of M(G) are also equal to M(G). Finally, we show this result implies M(G) is indeed a sigma-algebra.

Let <math>G</math> be an [[fieldField of sets|algebra of sets]] and define <math>M(G)</math> to be the smallest monotone class containing <math>G.</math> Then <math>M(G)</math> is precisely the [[Sigma-algebra|σ{{sigma}}-algebra]] generated by <math>G,</math>; i.e.that σis <math>\sigma(G) = M(G).</math>

}}

==Monotone class theorem for functions==

Let <math>\mathcal{B} = M(G)</math>, i.e. <math>\mathcal{B}</math> is the smallest monotone class containing <math>G</math>. For each set <math>B</math>, denote <math>\mathcal{B}_B</math> to be the collection of sets <math>A \in \mathcal{B}</math> such that <math>A \cap B \in \mathcal{B}</math>. It is plain to see that <math>\mathcal{B}_B</math> is closed under increasing limits and differences.

{{math theorem|name=Monotone class theorem for functions|note=|style=|math_statement=

Consider <math>B \in G</math>. For each <math>C \in G</math>, <math> B \cap C \in G \subset \mathcal{B}</math>, hence <math>C \in \mathcal{B}_B </math> so <math>G \subset \mathcal{B}_B </math>. This yields <math>\mathcal{B}_B = \mathcal{B}</math> when <math> B \in G</math>, since <math>\mathcal{B}_B</math> is a monotone class containing <math>G</math>, <math>\mathcal{B}_B \subset \mathcal{B}</math> and <math>\mathcal{B}</math> is the smallest monotone class containing <math>G</math>

Let <math>\mathcal{A}</math> be a [[Pi system|π{{pi}}-system]] that contains <math>\Omega\,</math> and let <math>\mathcal{H}</math> be a collection of functions from <math>\Omega</math> to '''<math>\R'''</math> with the following properties:

Now,# more generally, supposeIf <math>BA \in \mathcal{BA}</math>. For eachthen <math>C \in G</math>, we have <math> B \in \mathcalmathbf{B1}_C</math> and by the last result, <math>B \cap C_A \in \mathcal{BH}</math>. Hence,where <math>C \in \mathcalmathbf{B1}_B_A</math> sodenotes <math>Gthe \subset \mathcal{B}_B</math>,[[indicator andfunction]] soof <math>\mathcal{B}_B = \mathcal{B}A.</math> for all <math>B \in \mathcal{B}</math> by the argument in the paragraph directly above.

(2)# If <math>f, g \in \mathcal{H}</math>, and <math>c \in \Reals</math> then <math>f + g</math> and <math>cfc f \in \mathcal{H}</math> for any real number <math>c.</math>

# If <math>f_n \in \mathcal{H}</math> is a sequence of non-negative functions that increase to a bounded function <math>f</math> then <math>f \in \mathcal{H}.</math>

Then <math>\mathcal{H}</math> contains all bounded functions that are measurable with respect to <math>\sigma(\mathcal{A}),</math>, which is the {{sigma}}-algebra generated by <math>\mathcal{A}.</math>

Since <math>\mathcal{B}_B = \mathcal{B}</math> for all <math>B \in \mathcal{B}</math>, it must be that <math>\mathcal{B}</math> is closed under finite intersections. Furthermore, <math>\mathcal{B}</math> is closed by differences, so it is also closed under complements. Since <math>\mathcal{B}</math> is closed under increasing limits as well, it is a sigma-algebra. Since every sigma-algebra is a monotone class, <math>\mathcal{B} = \sigma\,(G)</math>, i.e. <math>\mathcal{B}</math> is the smallest sigma-algebra containing G

}}

===Proof===

==Monotone class theorem for functions==

The following argument originates in [[Rick Durrett]]'s Probability: Theory and Examples.<ref name="Durrett">{{cite book|last=Durrett|first=Rick|year=2010|title=Probability: Theory and Examples|url=https://archive.org/details/probabilitytheor00rdur|url-access=limited|edition=4th|publisher=Cambridge University Press|page=100[https://archive.org/details/probabilitytheor00rdur/page/n287 276]|isbn=978-0521765398}}</ref>

===Statement===

Let <math>\mathcal{A}</math> be a [[Pi system|π-system]] that contains <math>\Omega\,</math> and let <math>\mathcal{H}</math> be a collection of functions from <math>\Omega</math> to '''R''' with the following properties:

{{math proof|drop=hidden|proof=

(1) If <math>A \in \mathcal{A}</math>, then <math>\mathbf{1}_{A} \in \mathcal{H}</math>

(1)The Ifassumption <math>A\Omega\, \in \mathcal{A},</math> (2), thenand (3) imply that <math>\mathcal{G} = \left\{A : \mathbf{1}_{A} \in \mathcal{H}\right\}</math> is a {{lambda}}-system.

By (1) and the [[Dynkin system|{{pi}}−{{lambda}} theorem]], <math>\sigma(\mathcal{A}) \subseteq \mathcal{G}.</math>

The assumption <math>\Omega\, \in \mathcal{A}</math>,Statement (2) and (3) implyimplies that <math>\mathcal{G} = \{A: \mathbf{1}_{A} \in \mathcal{H}\}</math> is a λ-system. By (1) and the [[Dynkin system|π − λ theorem]], <math>\sigma(\mathcal{A}) \subset \mathcal{G}</math>. (2) implies <math>\mathcal{H}</math> contains all simple functions, and then (3) implies that <math>\mathcal{H}</math> contains all bounded functions measurable with respect to <math>\sigma(\mathcal{A}).</math>

}}

==Results and Applicationsapplications==

(2) If <math>f,g \in \mathcal{H}</math>, then <math>f+g</math> and <math>cf \in \mathcal{H}</math> for any real number <math>c</math>

(3)As Ifa corollary, if <math>f_n \in \mathcal{H}G</math> is a sequence[[ring of non-negativesets]], functionsthen thatthe increasesmallest tomonotone aclass boundedcontaining functionit <math>f</math>,coincides thenwith the [[Sigma-ring|{{sigma}}-ring]] of <math>f \in \mathcal{H}G.</math>

By invoking this theorem, one can use monotone classes to help verify that a certain collection of subsets is a [[Sigma-algebra|{{sigma}}-algebra]].

Then <math>\mathcal{H}</math> contains all bounded functions that are measurable with respect to <math>\sigma(\mathcal{A})</math>, the sigma-algebra generated by <math>\mathcal{A}</math>

The monotone class theorem for functions can be a powerful tool that allows statements about particularly simple classes of functions to be generalized to arbitrary bounded and measurable functions.

===Proof===

The following argument originates in [[Rick Durrett]]'s Probability: Theory and Examples.

<ref name="Durrett">{{cite book|last=Durrett|first=Rick|year=2010|title=Probability: Theory and Examples|edition=4th|publisher=Cambridge University Press|page=100|isbn=978-0521765398}}</ref>

==See also==

The assumption <math>\Omega\, \in \mathcal{A}</math>, (2) and (3) imply that <math>\mathcal{G} = \{A: \mathbf{1}_{A} \in \mathcal{H}\}</math> is a λ-system. By (1) and the [[Dynkin system|π − λ theorem]], <math>\sigma(\mathcal{A}) \subset \mathcal{G}</math>. (2) implies <math>\mathcal{H}</math> contains all simple functions, and then (3) implies that <math>\mathcal{H}</math> contains all bounded functions measurable with respect to <math>\sigma(\mathcal{A})</math>

* {{annotated link|Dynkin system}}

==Results and Applications==

* {{annotated link|π-λ theorem|{{pi}}-{{lambda}} theorem}}

As a corollary, if G is a [[Ring of sets|ring]] of sets, then the smallest monotone class containing it coincides with the sigma-ring of G.

* {{annotated link|Pi-system|{{pi}}-system}}

* {{annotated link|σ-algebra}}

==Citations==

By invoking this theorem, one can use monotone classes to help verify that a certain collection of subsets is a sigma-algebra.

{{reflist|group=note}}

The monotone class theorem for functions can be a powerful tool that allows statements about particularly simple classes of functions to be generalized to arbitrary bounded and measurable functions.

{{reflist}}

==References==

<references/>

* {{Durrett Probability Theory and Examples 5th Edition}} <!-- {{sfn|Durrett|2019|p=}} -->

==See also==

This article was advanced during a Wikipedia course held at Duke University, which can be found here: [[Education Program:Duke University/WIKIPEDIA AND ITS ANCESTORS: Rethinking Encyclopedic Knowledge in the Digital Age (Spring 2013)|Wikipedia and Its Ancestors]]

[[Category:SetFamilies familiesof sets]]

[[Category:Theorems in measure theory]]