Combining semantic networks with multi-attribute utility models: An evaluative data base indexing method

摘要

This article presents a method for representing evaluative knowledge by combining techniques from two separate but complementary disciplines. A semantic network knowledge representation is merged with a decision-theoretic multiattribute utility (MAU) model to provide an indexing technique for data bases that store information in the form of judgments or evaluations. Such a merger provides reasoning and inference capabilities that allow the data base to estimate an evaluation even if no direct evidence is available for user queries. In addition to the technique, a sample application in international marketing, called the Country Consultant, is described.