Equation (9.0)

Overview & Context

The Function Node Editor allows you to define what calculation or “function” occurs in the corresponding Function Node. As opposed to standard nodes, which feature states with probabilities, a Function Node produces a value as an output (numerical, boolean, or a string of characters). As such, it can use Inference Functions to calculate various measures of specific nodes or the network as a whole.

New Feature: Inference Function — ObsCount(h, l, p, n)

This new Inference Function counts the number of the specified types of evidence that are currently set in your Bayesian network:

• Hard Evidence (h)
• Likelihood Evidence (l)
• Probabilistic Evidence (p)
• Numerical Evidence (n)

You can define which types of evidence to count by setting the boolean parameters (TRUE, FALSE) for h, l, p, and n.

To perform this particular type of inference, the Function Node does not need to be connected to any node in your network.