Introduction

To this end, we introduce partitioning classes that occupy higher positions in the ontology hierarchy, serving as superclasses or overarching categories that help structure the ontology. They can define broad divisions or partitions within the domain, with extensions of regular classes positioned within these partitions based on their properties or relationships. Regular classes, on the other hand, can exist at different levels of the ontology hierarchy without necessarily serving as partitioning elements.

The key research question is: How can we design a methodology that provides a unified and standardized view of knowledge partitioning? This raises other questions: What types of knowledge domain partitioning currently exist and what are their limitations? How do these limitations affect conceptual modeling and to what extent do they prevent the implementation of more powerful search, retrieval and inference methods?

The rest of this paragraph is organized as follows: In the Related Work, we discuss what approaches already exist for the partitioning of classes and what shortcomings they have. Basic naming conventions and formal definitions are then introduced in the Preliminaries. The Methodology introduces Partitioning Classes (PC) using examples for attributes such as ‘gender’, ‘color’, ‘warm-blooded’, and ‘LifePhase’. As a special type of PCs the concept of Boolean PCs is also described. We also explain, how the properties of PCs, such as multiplicity und mutability are preserved. In the Evaluation, the question of whether it is possible to save storage space and to increase performance when processing conjunctive queries is examined. In the Results we use the Multi-Level Modeling (MLM) of other authors to compare how PCs can be used to streamline conceptual modeling. Under the heading of orthogonality, we also discuss how more flexibility and richness in modeling can be gained through the use of PCs. In the Results, we summarize the results.