First you have prior probability say 50% . When observation is made we use that to update our prior probability = posterior probability 66%. If another observation is made our prior probability 66% which we would update with new observation right? and we continue this process?

I just want to clarify if my understanding it correct.

  • 2
    $\begingroup$ Yes. Iterative Bayesian updating is exactly that. $\endgroup$ – usεr11852 May 22 '20 at 0:18

Just one thing, prior is a probability distribution, not a value like 50% (prior may be also improper.)

As an example with the Binomial and Beta prior:

Starting with a prior distribution $Be(1,1)$ (this is the uniform distribution).

We see 5 outcomes, where 3 are positive and 2 negative.

The posterior is now $Be(1+3,1+2)$ = $Be(4,3)$. This is the prior for the next time.

We see another 13 outcomes, 8 positive, 5 negative.

The posterior is now $Be(4+8,3+5)$ = $Be(12,8)$.

We see 125 outcomes, 80 positive, 45 negative.

The posterior is now $Be(12+80,8+45)$=$Be(92,53)$.

Here is each curve.

enter image description here


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.