0
$\begingroup$

I'm going over markov models and I'm tying to understand the following notation for chain probability:

$$p(a_1, a_2, a_3, a_4) = p(a_1)p(a_2\mid a_1)p(a_3\mid a_1,a_2)p(a_4\mid a_1,a_2,a_3)$$

"The Chain Rule is one simple consequence of the definition of conditional probability: the joint probability of some set of events $a_1, a_2, a_3, a_4$ can also be expressed as a ‘chain’ of conditional probabilities"

Can someone explain?

$\endgroup$
  • $\begingroup$ Unrelated to the Markov property, always true, simply iterates Bayes formula $$p(a,b)=p(a)p(b\mid a)$$ for various $a$ and $b$. $\endgroup$ – Did Mar 9 '17 at 11:20
3
$\begingroup$

By definition of conditional probability, we have $P(A,B) = P(A\mid B)P(B)$, where $P(A,B)$ means $P(A\cap B)$.

Now taking $A=a_4$ and $B = (a_1,a_2,a_3)$, we get: $$P(a_4,a_3,a_2,a_1) = P(a_4\mid a_1, a_2, a_3)P(a_1,a_2,a_3)$$ Now with $A = a_3$ and $B=(a_1,a_2)$ the last factor can be written $$P(a_1,a_2,a_3) = P(a_3\mid a_1,a_2)P(a_1,a_2)$$ Once again, $P(a_1,a_2)$ is $P(a2\mid a_1)P(a_1)$ and you get your formula.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.