huge_network subset (N=25)¶
The first 25 nodes of data/huge_network.json in topological order. Taking a topological prefix preserves parent-child structure: every retained node's parents are also in the subset, so the result is a well-formed Bayesian network from the same topology family as the full 39-node fixture.
Bayesian network (DAG)¶
flowchart TB
node11["node11"]
node12["node12"]
node14["node14"]
node15["node15"]
node16["node16"]
node19["node19"]
node20["node20"]
node21["node21"]
node22["node22"]
node23["node23"]
node26["node26"]
node27["node27"]
node28["node28"]
node29["node29"]
node3["node3"]
node30["node30"]
node39["node39"]
node4["node4"]
node5["node5"]
node6["node6"]
node7["node7"]
node8["node8"]
node9["node9"]
node10["node10"]
node13["node13"]
node11 --> node10
node12 --> node10
node14 --> node13
node15 --> node13
Junction forest¶
The junction graph here is a forest of 21 disconnected components, not a single tree. Cliques are labelled with their variable sets; edges labelled with the separator (the intersection of the two cliques' domains). Cliques without an edge to anything are isolated components.
flowchart LR
C0["C0<br/>node3"]
C1["C1<br/>node4"]
C2["C2<br/>node5"]
C3["C3<br/>node6"]
C4["C4<br/>node7"]
C5["C5<br/>node8"]
C6["C6<br/>node9"]
C7["C7<br/>node10, node11, node12"]
C8["C8<br/>node13, node14, node15"]
C9["C9<br/>node16"]
C10["C10<br/>node19"]
C11["C11<br/>node20"]
C12["C12<br/>node21"]
C13["C13<br/>node22"]
C14["C14<br/>node23"]
C15["C15<br/>node26"]
C16["C16<br/>node27"]
C17["C17<br/>node28"]
C18["C18<br/>node29"]
C19["C19<br/>node30"]
C20["C20<br/>node39"]