The Key Concepts Extraction (KCE) algorithm defines a novel approach to identifying the concepts in an ontology, which best summarize what the ontology is about. KCE makes use of a number of criteria, drawn from cognitive science, network topology, and lexical statistics to try and produce the kind of ontology summaries which human experts would come up with. Indeed a formal evaluation of the method, described in Peroni et al. (2008), has shown an excellent degree of correlation between the ontology summaries produced by KCE and those produced by human experts. While the generation of automatic methods for ontology summarization is an interesting research area in itself, KCE also provides a basis for novel approaches to a variety of ontology engineering tasks, including ontology matching, automatic classification, ontology modularization, and ontology.
The Key-Concepts Extraction Service for OWL ontologies is an online service, which automatically extracts the most representative classes of an ontology. In the current version, the service identifies key concepts in an ontology by combining the notion of natural categories, which aims to identify concepts that are information-rich in a psycho-linguistic sense and the notion of density, which highlights concepts which are information-rich in an ontological sense. In addition, it also uses a coverage criterion, to ensure that no important part of the ontology is ignored in the resulting selection. More details about KCE can be found in Peroni et al. (2008).
The following pseudo-URL describes the way to call the service:
is the URL to call the service.
is the full http://... URL of the OWL ontology that will be processed by the service. It must be always the last item in the pseudo-URL, and may be preceded by one or more (slash-separated) parameters.
optional slash-separated parameters:
When this optional parameter is specified, the axioms of the imported ontologies of ontology-url are considered for the computation.
The result will be returned in HTML.
The result will be returned in XML.
The result will be returned in CSV (it is the default).
When this optional parameter is specified, the specified number of key concept will be extracted (default: 20).
KCE is available under an open source license and the sources can be accessed at http://sourceforge.net/projects/kce/.