Sigma Algebra
Sigma-Algebra (Borel Field)
Let
is closed under complement: - If
, then
- If
is closed under countable unions:- If
is a countable collection of elements of and , then
- If
Additionally, is follows that:
(empty set)- Since (1) says
, and (2) says , but
- Since (1) says
- If
is a countable collection of elements of and , then- "
is closed under countable intersections"
- "
In the context of a probability space, we say that
If
However,
There are two extreme examples of a sigma algebra:
- The collection
- The power set of
,
Any sigma algebra
Credit: some stack exchange answer.
Focusing specifically on the application to probability:
The first property states that we always know the probability of something happening (1) and nothing happening (0)
For the second property, suppose we have a coin flip with
For the third property, suppose we roll a die, so
This may seem like the event space. In fact, the event space is a
Atom
An atom of a
In other words, an atom of a
Consider the set
Partition forms Sigma Algebra
See partition
The collection
Consider the set
Then the sigma algebra formed by these partitions is:
Generated Sigma Algebras
Let
In other words, the sigma algebra generated by
Given
The first two sets satisfy are requirements of a
The next set (
The next set (
The final set (
Finally, this is the smallest
Existence of a Generated Sigma Algebra
If
If you take the intersection of all
Dynkin's π-λ Theorem
Π-System
A collection
Λ-System (Dynkin System)
A Λ-System is Closed Under Differences
See set difference
If
In the first case, assume
Recall from the definition of a set difference that
Note that
Additionally, since
Since
Since
In the second case, Assume
A ΠΛ-System is a Σ-Algebra
A family which is both a
Let
Given the definitions of a
Let
We can re-write
where
so
Clearly,
Since
Since
The Theorem
If
That is, the sigma algebra generated by
Proof of the theorem (and additional lemmas): https://www.math.lsu.edu/~sengupta/7312s02/sigmaalg.pdf