Calculating attribute values using inheritance structures in fuzzy object-oriented data models

The problem of dealing with incomplete information in object-oriented data models (OODMs) is addressed. A method to compute default values for unknown objects' attributes is proposed, based both on the association of typical values with the attributes in the intensional definition of a class and on the application of a prioritized aggregation operator to combine typical values appearing in an inheritance structure. This method can also be applied to refine vague attribute values expressed by means of fuzzy sets interpreted as possibility distributions. A new interpretation of partial inheritance in this context is proposed, introducing the concept of "partial overriding" of typical values.