A category theory approach to conceptual data modeling

Abstract

This paper describes a category theory semantics for conceptual data modeling. The conceptual data modeling technique used can be seen as a generalization of most existing conceptual data modeling techniques. It contains features such as specialization, generalization, and power types. The semantics uses only simple category theory constructs such as (co)limits and epi- and monomorphisms. Therefore, the semantics can be applied to a wide range of instance categories, it is not restricted to topot or cartesian closed categories. By choosing appropriate instance categories, features such as missing values, multi-valued relations, and uncertainty can be added to conceptual data models.