The semantics of parts versus aggregates in data/knowledge modelling

Abstract

The incorporation of semantics into conceptual models has for long been a goal of the data/knowledge modelling communities. Equally, conceptual models strive for a high degree of intuitiveness in order be better understood by their human users. This paper aims to go one step in this direction by introducing the parl-of relation as a special case of aggregation. To do so we investigate the semantic constraints accompanying this specialization and suggest different ways of incorporating part-of semantics into data/knowledge models. Further, it is demonstrated that, in analogy with IS-A relations, part-of relations form hierarchies (dag’s) which constitute an important conceptual aid in understanding complex systems. Finally, we investigate the conditions under which the part-of relation exhibits transitive behavior which can be exploited for automated inferences facilitated by the transitivity property.