A function node represents a real-valued function that depends on (some or all of) the parents of the node. Function nodes are not (directly) involved in the inference process - evidence cannot be specified for function nodes, but the function associated with the node can be evaluated using the results of inference or simulation as input. Add a Function node by using the Function Tool.
Figure 1: A Function node (F1) with a discrete and a continuous parent.
Function node values are specified as an expression in the node model. A number of rules apply for Function Nodes: