4
$\begingroup$

Does anyone know of a generalization of Bayes' theorem to multiple conditions?
From this answer I can see the definition of conditional probability with multiple conditions, but I couldn't find any reference to a statement of Bayes' theorem with several conditions. I only need a generalization to 2 conditions for what I'm working on, but I'm also curious to its statement in $n$ conditions.

Perhaps if we had the multiplication rule, too, we could derive Bayes' rule from the above?

$\endgroup$

1 Answer 1

7
$\begingroup$

One generally accepted definition for $\mathbb{P}\left(B\right) \neq 0$ is $$\mathbb{P}\left(A \mid B\right) = \dfrac{\mathbb{P}\left(A \cap B\right)}{\mathbb{P}\left(B\right)}\text{.}$$ If we had $n$ conditions, we could suppose that $\bigcup\limits_{i=1}^{\infty}A_i = A$, where $A_1, A_2, \dots$ is a partition of an event $A$, and $$\mathbb{P}\left(A_i \mid B_1 \cap B_2 \cap \cdots\cap B_n\right) = \dfrac{\mathbb{P}\left(A_i \cap B_1 \cap B_2 \cdots \cap B_n\right)}{\mathbb{P}\left(B_1 \cap B_2 \cdots \cap B_n\right)} = \dfrac{\mathbb{P}\left(B_1 \cap B_2 \cap \cdots\cap B_n \mid A_i\right)\mathbb{P}\left(A_i\right)}{\sum\limits_{A_i \subseteq A}\mathbb{P}\left(B_1 \cap B_2 \cap \cdots\cap B_n \mid A_i\right)\mathbb{P}\left(A_i\right)}\text{.}$$ [For someone who knows this material better than I do, feel free to let me know if I'm wrong somewhere. I'm basing this on intuition.] The formula above is particularly useful for introductory Bayesian methods (finding posterior distributions, in particular).

$\endgroup$
1
  • $\begingroup$ Here you’ve extended A and B to be sets of multiple events. I think the OP was specifically asking about the events on which A is conditioned (i.e. extending B only). At first, I was confused when you introduced the set of A’s $\endgroup$
    – Earlien
    Jan 12, 2020 at 22:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .