How would be the right method to use an Upper or Core Ontology. I have developed a Upper Ontology to be used in the Manufacturing Domain. My intention is to create domain ontologies on demand and link them to the Upper Ontology via subclasses and subproperties. Would that be the way to go? I would appreciate any opinions, suggestions of design patterns, link to papers, etc.
Hi Irlan!
We have been proposing the Building Topology Ontology (BOT) and include alignments to both upper and more specific domain ontologies (http://www.semantic-web-journal.net/content/bot-building-topology-ontology-w3c-linked-building-data-group-0). The extensions/ linking has been formalised through specialisation (rdfs:subClassOf, rdfs:subPropertyOf). The Approach has ben Chosen as we felt logically it seems to be a good fit.
Still I would be also interested into additional comments by ontology experts how to extend upper ontologies into specific Domains.
Best
Georg
You can go for a multi-level modelling approach. This way an upper level ontology can be on one level and then the specific domain ontologies on a lower level. However, if depends if the lower level ontologies are mere specialisations or partial instantiations. You can look for multilevel approaches like PowerType modelling or Deep instantiation.
We have used stereotypes for enriching terms of a Top-Domain Software Testing Ontology named TestTDO (see "TestTDO: A Top-Domain Software Testing Ontology" paper available at https://www.researchgate.net/publication/339313649_TestTDO_A_Top-Domain_Software_Testing_Ontology)
TestTDO stereotyped terms are enriched with terms of higher-level ontologies at Core and Foundational levels (e.g. ProcessCO, SituationCO, ThingFO). We have developed FCD-OntoArch, which is a four-layered ontological architecture that considers foundational, core, domain and instance levels.
From my point of view, this is due to the modularization concept on which the ontology is based and its layered structure. According to our experience, one should not only look at the technological aspects but also at the underlying concept and structure of the ontology. We have addressed these aspects in the following paper: Article Digitalisierung im Engineering - Ein Ansatz für ein Vorgehen...
In addition, the question arises for me as to what this ontology should be used for? In our case, we need the ontologies or data model for knowledge-based configurators, which should serve a direct automated use of knowledge. Here we proceeded as follows: Article A method for collaborative knowledge acquisition and modelin...
I have developed an Upper Ontology as well as some domain ontologies. I'm searching for pros and cons of the following.
Dear Irlan@
Answer for Q#1: We mainly enrich DO terms by means of stereotypes of CO terms. In turn these CO -as ProcessCO, SituationCO- terms are enriched by ThingFO terms.
ProcessCO available at DOI: 10.13140/RG.2.2.31158.32328
SituationFO available at DOI: 10.13140/RG.2.2.11359.82089/3
ThingFO available at DOI: 10.13140/RG.2.2.11163.21281/1
Example of a top DO enriched with the above Ontolgies is
TestTDO available at DOI: 10.13140/RG.2.2.26338.07368/1
Answer for Q#2: A Particular Thing (e.g. Task, Testing Activity, Incident, Student, etc.) may create instances. See the four-layered Architecture in those documents.
Sincerely,
Luis
- Badges
- Science topic
- Similar topics
- Papers