GRAPH-GRAMMAR ASSISTANCE FOR AUTOMATED GENERATION OF INFLUENCE DIAGRAMS

One of the most difficult aspects of modeling complex dilemmas in deci sion-analytic terms is composing a diagram of relevance relations from a set of domain concepts. Decision models in many domains, however, e xhibit certain prototypical patterns that can guide the modeling proce ss. Concepts can be classified according to semantic types that have c haracteristic positions and typical roles in an influence-diagram mode l. We have developed a graph-grammar production system that uses such inherent interrelationships among terms to facilitate the modeling of medical decisions. Our system also can examine a set of graph-grammar rules to establish whether the grammar satisfies a number of propertie s that we have determined to be important in the derivation of influen ce-diagram models. Our findings suggest that syntactic patterns can le ad to automated construction of decision models in domains other than medicine.