Types of set algebras

Measure theory perspective

In this table the class of sets $\mathcal{A}$ is a collection of subsets from another set $\Omega$

System Type	Definition	Deductions
Ring	$\mathcal{A}$ is $\backslash$-closed $\mathcal{A}$ is $\cup$-closed	$\emptyset\in\mathcal{A}$
$\sigma$-ring	$\mathcal{A}$ is a ring $\mathcal{A}$ is $\sigma$-$\cup$-closed	$\sigma$-$\cap$-closed also
Algebra	$\mathcal{A}$ is closed under complements $\mathcal{A}$ is $\cup$-closed	$\mathcal{A}$ is $\backslash$-closed $\Omega\in\mathcal{A}$
$\sigma$-algebra	$\mathcal{A}$ is an algebra $\mathcal{A}$ is $\sigma$-$\cup$-closed	also $\sigma$-$\cap$-closed
Semiring	TODO: Page 3 in
Dynkin system

These types are all related and I have a nice diagram to remember this which uses arrow directions to 'encode' the difference. In my diagram upwards arrows indicate something to do with union, with $\cup$, downwards with $\cap$ (think "make bigger"=up, which is union and "going down" being cap. A rightward slant means "sigma-whatever-the-vertical-direction-is" which means closed under countable whatever. Lastly, a horizontal arrow indicates membership, right means "contains entire set" and that's all that is used. Lastly:

• All paths lead to $\sigma$-algebra

$$\begin{xy}\xymatrix{ & & \text{Dynkin system} \ar[d]^{\cap\text{-closed}} \\ & {\sigma\text{-ring}} \ar[r]^{\Omega\in\mathcal{A}} & {\sigma\text{-algebra}} \\ \text{ring} \ar[ur]^(.4){\sigma\text{-}\cup} \ar[r]^{\Omega\in\mathcal{A}} & \text{algebra} \ar[ur]^(.4){\sigma\text{-}\cup} & \\ \text{semiring} \ar[u]_{\cup\text{-closed}} & & }\end{xy}$$
Alec's 'super' diagram

Notice in addition the nice symmetry of the diagram (the line of symmetry would be from top left to bottom right), it doesn't preserve arrow directions, and obviously not names, but shape.

Overall this is a very easy diagram to remember. I remember ring easily (it's what you'd need to "do probability" on, unions and set-subtractions, and the empty set (required to have subtractions anyway)). This lets me build the rest. The only not obvious ones are Dynkin-systems and semirings

Relationship between all types

This of course isn't the entire picture. In addition we can use the Borel $\sigma$-algebra on a topology to get a $\sigma$-algebra, the below diagram is more complete, at the cost of ease to remember

$$\begin{xy}\xymatrix{ & & \text{Dynkin system} \ar[d]^{\cap\text{-closed}} & \\ & {\sigma\text{-ring}} \ar[r]^{\Omega\in\mathcal{A}} & {\sigma\text{-algebra}} & \\ \text{ring} \ar[ur]^(.4){\sigma\text{-}\cup} \ar[r]^{\Omega\in\mathcal{A}} & \text{algebra} \ar[ur]^(.4){\sigma\text{-}\cup} & & \text{topology}\ar@{.>}[ul]_(.25){\text{Borel }\sigma\text{-algebra}} \\ \text{semiring} \ar[u]_{\cup\text{-closed}} & & & }\end{xy}$$

Diagram showing ALL the relationships

Other Notes

Closed under
Type	$\sigma\in\mathcal{A}$	$\bigcap$	$\sigma$-$\bigcap$	$\bigcup$	$\sigma$-$\bigcup$	$-$	$C$
Semi-Ring
Ring
$\sigma$-Ring
Algebra
Dynkin system
$\sigma$-Algebra	#	#		X	X	#