A decision tree network is a decision network in which every predecessor of a decision node is a <em>direct</em> predecessor of that node. A recommendation for action can be obtained from a consistently assessed decision tree network without applying any reversal operations (ie, without applying Bayes’ theorem). A decision tree network can be directly converted into a decision tree.

See also: Addition, Evaluating a Relevance Diagram, Removal, Well-Formed Decision Relevance Diagram, and Well-Formed Partial Decision Relevance Diagram.