Static Classification Schemes for an Object System

May 1998

Proceedings of the Eleventh International Florida Artificial Intelligence Research Society Conference (FLAIRS 1998)

Volume

Proceedings of the Eleventh International Florida Artificial Intelligence Research Society Conference (FLAIRS 1998)

All Papers

Instance (or object) classification in a knowledge base management system is the very same problem as view determination in an object DBMS, where views are subsets of classes intensionally defined by constraints. Maintaining the relationship between an object and its current views when the object evolves is a difficult and costly problem. We propose a solution for instance classification based on a partitioning of the object space obtained through a static analysis of the properties defining the viev, s. Four kinds of properties are considered: domain constraints, interattribute dependencies, dynamic (past temporal) constraints, and linear constraints.