** Next:** QUASI-BAYESIAN THEORY AND POLYTOPIC
**Up:** LOCAL ROBUSTNESS ANALYSIS OF
** Previous:** LOCAL ROBUSTNESS ANALYSIS OF

We consider a set x of discrete variables; each variable
x_{i} has a finite set of values x_{i} and a set of variables
pa(x_{i}), the *parents* of x_{i}. A Bayesian network
defines a unique joint probability distribution [29]:

p(x) = prod_{i} p(x_{i} | pa(x_{i})).

We use the abbreviation p_{i} for
p(x_{i}|pa(x_{i})); expression (1) can be
written as p(x) = prod_{i} p_{i}.
Suppose a set of variables is fixed as evidence e;
p^{e}() is a distribution where variables e are fixed.
Our algorithms assume efficient computation of posterior marginals in
a Bayesian network [12, 21, 40].

© Fabio Cozman[Send Mail?]

Fri May 30 15:55:18 EDT 1997