Proof: Let be a positive distribution defined by a
Markov network where are the adjacencies of ,
. Construct a consistent dependency network from by
extracting the conditional distributions
. Because these probabilities came from the Markov
network, we know that
so that the adjacencies in the dependency-network
structure are the same as those in the Markov network.
Now let be a positive distribution encoded by a consistent
dependency network. By definition, we have that is independent
of
given , . Because is
positive and these independencies comprise the global Markov property
of a Markov network with
, , the
Hammersley-Clifford theorem (Besag, 1974; Lauritzen,
Dawid, Larsen, and Leimer, 1990) implies that can
be represented by this Markov network.
Theorem 2: An ordered Gibbs sampler applied to a
consistent dependency network for , where each is finite
(and hence discrete) and each local distribution
is
positive, defines a Markov chain with a unique stationary joint
distribution for equal to that can be reached from any
initial state of the chain.
Proof: In the body of the paper, we showed that the
Markov chain can be described by the transition matrix
, where is the ``local'' transition
matrix describing the resampling of according to the local
distribution
. We also showed that this Markov chain has
a unique joint distribution that can be reached from any starting
point. Here, we show that is that stationary
distribution--that is,
,
where
. To do so, we
show that for each ,
.
Theorem 4: A minimal consistent dependency network for
a positive distribution must be bi-directional.
Proof: We use the graphoid axioms of Pearl
(1988). Suppose the theorem is false. Then, there
exists nodes and such that is a parent of and
is not a parent of . Let
,
, and
. From
minimality, we know that
does not
hold. By decomposition,
does not
hold. Given positivity, the intersection property holds. By the
intersection property, at least one of the following conditions does
not hold: (1)
, (2)
. (If
, then
condition 2 holds vacuously.)
Now, from Theorem 2, we know that
. Using weak union, we have that
--that is, condition 1 holds. Also,
from Theorem 2, we know that
--that
is, condition 2 holds, yielding a contradiction.