Models

Conditional probability tables (CPTs) for discrete chance nodes, utility tables for utility nodes, and initial policies for decision nodes can be specified compactly through models. A model consists of a list of model nodes and a set of Expressions (one expression for each configuration of the states of the model nodes). The model nodes constitute a subset of the parents of the node to which the model belongs. This subset is specified under the Table Tab of the Node Properties dialog box.

Specification of CPTs through model is particularly useful when the conditional probability distribution for a variable is complex and follows (at least approximately) certain functional or distributional forms. In such cases it is cumbersome to specify the conditional probability table (CPT) manually.

When one needs different expressions depending on the states of one or more parent nodes, the use of model nodes can be quite convenient, as complicated expressions can be considerably simplified.

If there are no model nodes, the model contains a single expression.

Examples of the use of models and model nodes are given in the Table Generator Tutorial.