To be able to use nodes as arguments in expression functions, node types are introduced for discrete chance nodes and decision nodes. Below are listed the different node types:
Labelled - Each state has a label. This is similar how all discrete nodes looked in old versions of HUGIN.
Boolean - Has two states: false and true.
Numbered - Each state has a numeric value. This means that you can make expressions define the sum of two numbered nodes. You can also specify standard discrete distribution functions for numbered nodes if you fit the state values to what is required for the distribution function (typically: 0, 1, 2,…).
Interval - Each state represents an interval on the real axis. The entire set of states span a continuous interval on the real axis. This is used to approximate a continuous random variable. You can specify standard continuous distribution functions for interval nodes, or you can use them in arithmetic functions.