Encyclopedia_Junction Tree

Netica

Junction Tree

 Also known as a “join tree”.

A junction tree is the internal structure that Netica uses for = 4 && typeof(BSPSPopupOnMouseOver) == 'function') BSPSPopupOnMouseOver(event);" class="BSSCPopup" onclick="BSSCPopup('X_PU_belief_updating.htm');return false;">belief updating.  Netica compiles a = 4 && typeof(BSPSPopupOnMouseOver) == 'function') BSPSPopupOnMouseOver(event);" class="BSSCPopup" onclick="BSSCPopup('X_PU_Bayes_net.htm');return false;">Bayes net or = 4 && typeof(BSPSPopupOnMouseOver) == 'function') BSPSPopupOnMouseOver(event);" class="BSSCPopup" onclick="BSSCPopup('X_PU_decision_nets.htm');return false;">decision net into a junction tree for efficiency.

The junction tree T of = 4 && typeof(BSPSPopupOnMouseOver) == 'function') BSPSPopupOnMouseOver(event);" class="BSSCPopup" onclick="BSSCPopup('X_PU_triangulated.htm');return false;">triangulated net G is a tree with the = 4 && typeof(BSPSPopupOnMouseOver) == 'function') BSPSPopupOnMouseOver(event);" class="BSSCPopup" onclick="BSSCPopup('X_PU_clique.htm');return false;">cliques of G as nodes, such that for every node N of G, if we remove from T all cliques not containing N, the remaining subtree remains connected.  In other words, any two cliques containing N are either adjacent in T or connected by a path made entirely of cliques that contain N.

You can examine the junction tree that Netica creates for a Bayes net by making a text report using Report Junction Tree after compiling.