Representation of WHAT-Knowledge Structures as Ontology Design Patterns

Considering any complex object several different aspects and layers can be investigated. WHAT-knowledge is one of the question-based aspects of knowledge that implies a conceptual representation of knowledge and is a major priority in learning. WHAT-knowledge key concepts are Entity, Concept, Class, Instance, Property. WHAT-knowledge key relationships are subClassOf, hasPart/isPartOf, type, classifies/isClassifiedBy. A classical way of teaching assumes that the studied concept definition is given first, then the concept properties are studied, and only after that - the relations with other concepts. To improve semantic interoperability, knowledge can be represented using description logics, however, it is a non-trivial task. As a consequence, created ontologies are often not scientific, i.e. they contain disproportionate concepts, tautologies, concepts with undefined disjoint relations, and so on. To provide an accurate representation of WHAT-knowledge, we proposed to represent WHAT-Knowledge Structures as Ontology Design Patterns. Typical WHAT-knowledge structures are the follows: determinables, contrast sets, genus/differentiate, taxonomies, faceted taxonomies, cluster concepts, family resemblances, graded concepts, frames, definitions, rules, rules with exceptions, essence and state assertions, opposites and contraries, relevance, and so on. To represent these structures we developed the following patterns: concept existence, incomparable concepts, concept properties, comparable concepts (contrast set (ordered and disordered), individual as an example of the class, intensional definition of concept, extensional definition of concept (implicational definition of concept, definitional concept description, rule-like definition of concept), concept exclusion, concepts intersection). Also, we proposed to use the following roles (non-taxonomic relations) that are defined as WHAT-Knowledge Structures through Ontology Design Patterns: existence role, existence role between the concepts, inverse roles, disjoint roles, roles inclusion, and chain of roles. The proposed approach allows increasing efficiency of the domain conceptual modeling, and model intelligibility for end-users. It can be used for the conceptual modeling and information retrieval in a wide range of domains and particularly in education and scientific research.