MathOverflow will be down for maintenance for approximately 3 hours, starting Monday evening (06/24/2013) at approximately 9:00 PM Eastern time (UTC-4).

## Inferring partial joint distribution given the entire joint distribution of a tree graphical model

Hi everyone

Now suppose I have a tree graphical model with joint probability density function $p(X_1,X_2,X_3,...,X_n)$ and I'd like to compute the joint density of $p(X_1,X_2,X_3)$. I know it's easy to use message passing or belief propagation to compute all the marginals $p(X_1)$, $p(X_2)$, $p(X_3)$, but it seems there are no efficient methods to compute the density of $p(X_1,X_2,X_3)$.

Does anybody have ideas?

Many thanks!!!

Rgds Hang

-