Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I want to combine two beta distributions to find the the posterior distribution as the precision-weighted combination of the prior and the likelihood distributions. As the prior and likelihood I have two beta distributions. How do i combine them? Want to know about all the theoretical techniques available. Is there any article which may help me in the analysis. Even guiding me with a weblink to any article would also help!

Regards, Bik

share|improve this question
add comment

1 Answer

up vote 1 down vote accepted

It is unclear exactly what you are asking for.

If your prior distribution is $B(\alpha,\beta)$ then the prior density is proportional to $x^{\alpha-1} (1-x)^{\beta-1}$.

Your likelihood function is perhaps proportional to $x^{\gamma} (1-x)^{\delta}$.

So you multiply these together and your posterior density is proportional to $x^{\alpha+\gamma-1} (1-x)^{\beta+\delta-1}$ which means your posterior distribution is $B(\alpha+\gamma,\beta+\delta)$.

share|improve this answer
    
Hello @Henry! Firstly, thanks a ton for helping me. Let me give more explanation on my query: I am trying to combine two beta distributions B1 (a1,b1) and B2 (a2,b2) using spreadsheet to get another alpha (a3) and beta (b3) for a combined beta distribution. Using this technique I intend to combine expert judgement and data to recalibrate my model. One of the option which I have is to take a weighted average of the paramaters (a1,b1,a2 and b2) to arrive at a3 and b3. Just wanted to inquire if there are other techniques available. Looking forwark for your response... –  Bik Feb 15 '13 at 13:57
    
What is supposed to be distributed with a Beta distribution? The parameter of another distribution? –  Henry Feb 15 '13 at 19:28
    
Yes... you are right... –  Bik Feb 16 '13 at 8:02
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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