Show 
Creating a Decision Net

Creating a decision net is pretty much the same as creating a Bayes net, so it involves adding the nodes and links, setting node properties and entering node tables.

One difference, obviously, is that you must designate some nodes as decision and utility nodes.  Usually you do this when you first add them, by using the image\RectTOOL.gif toolbar button to add decision nodes, and the image\HexagonTOOL.gif button to add utility nodes.  However, you can change node kinds using the node dialog box, or by selecting the nodes you wish to change, and then clicking on the toolbar button for the desired kind, while holding down the ctrl key.

You don’t enter a table for decision nodes, since Netica finds that when it solves the decision net.

Utility nodes must be continuous, and do not need to be discretized.  You may have multiple utility nodes, and Netica will use the sum of them as the total utility.  Utility nodes should not have children.

There should be at least one directed path through all the decision nodes to indicate their ordering in time.  You do not need to add no-forgetting links; Netica will add any necessary ones while solving the decision net.
 

Home > Decision Nets > Creating a Decision Net