A fuzzy concept network model and its applications

Computer systems often lack the ability to associate similar concepts. Compensating for the semantic ignorance of the system requires the modeling of the concept space of the application domain. A representation of the concept space as a network of fuzzy relationships is given. In the concept knowledge model, four types of relationships are distinguished and formally described. The transitivity of the relationships can be used to infer further relationships which have not been specified explicitly. T-norm operators are used to control the extent of transitivity. A mechanism is introduced which changes the behavior of the concept knowledge base according to the context of its use. Primary design goals are simplicity of the concept knowledge model and ease of construction of the concept knowledge base. An application of this concept knowledge which supports information access is presented