Ad-hoc and personal ontologies: a prototyping approach to ontology engineering

Large scale or common ontologies tend to be developed using structured and formal techniques that can be equated to the Waterfall system development life cycle. However, in domains that are not stable or well-understood a prototyping approach may be useful to allow exploration and communication of ideas. Alternatively, the ontology may be part of an intermediate step or representation that provides structure, organization, guidance and semantics for another task or representation. Given that the ontology is not the end goal and possibly not reusable, the overhead of developing or maintaining such ontologies needs to be minimal. This paper reviews some of the research using ad-hoc, one-off and, sometimes, throw away, personal ontologies and provides an example of a simple technique which uses Formal Concept Analysis to automatically generate an ontology as needed from a number of data sources including propositional rule bases, use cases, historical cases, text and web documents covering a range of applications and problem domains.