A Class is an optional property of a node. Classes are not exclusive, i.e., a node can be associated with more that one class. They are particularly useful to group nodes with common properties, and then, carry out actions on all the nodes that belong to the class.
Classes were first updated in versions 5.0.4, 5.0.7, 5.2, 5.4, and 6.0.
As of version 8.0, Depth is the only Predefined Class that can be automatically generated. Entirely based on the qualitative analysis of the graph, this new tool generates one class per depth/layer of the network. As Bayesian networks are Directed Acyclic Graphs, there are at least two depths (except for "networks" with one node). Whereas class [Depth_0] contains the root nodes (without parents), the last one, [Depth_n], contains the leaves node (without children).
These classes can be utilized, for example, for setting the layout of a graph.
Let's use the network CarStarts to illustrate this feature.
Five classes have been identified, with ten root nodes, and one leaf.