Representation and support generation in fuzzy relational databases

It is pointed out that standard relational database management systems are incapable of representing and manipulating the imprecise, incomplete, and vague information that is typically encountered in complex problem domains. Fuzzy relational databases have been developed to increase the flexibility in the representation of domain information. The different models of fuzzy relations and their roles in the framework of fuzzy relational database management systems are reviewed. The applicability of fuzzy relational database techniques as a methodology for approximate reasoning is examined. Many similarities exist in the processes that are used for fuzzy relational operations and those of approximate reasoning; however, the goals are often different. In fuzzy relational database theory, the objective is measuring similarity. For approximate reasoning, the objective is determining maximal compatibility. Several classes of similarity measures are examined to determine their appropriateness as compatibility measures in approximate reasoning