MAPLE (MAPping Architecture based on Linguistic Evidences) is an architecture and a software platform for the semi-automatic configuration of robust matching systems based on linguistic evidences and exploiting web published (linguistic) resources.
The proposed architecture is based metadata, since MAPLE suggests promising strategies by reasoning on metadata, describing ontologies, mediators and third-party (linguistic) resoureces published in the Web. MAPLE uses metadata about the input ontologies to figure out the type of mediation problem at hand, and to find a mediator capable of handling it.
The key assumption is that the linguistic grounding of the input ontologies is the primary source of evidences for reconciling them. Accordingly, MAPLE leverages the LIME vocabulary for characterizing the linguistic expressivity of the input ontologies, and for describing linguistic resources that may be used by mediators. LIME is also used in the description of ontology mediators, to express parametric dependencies on abstract linguistic resources, lately grounded on the characteristics of the ontologies to be matched.
MAPLE also assists the developers of mediation algorithms by providing interfaces for accessing the input ontologies in a uniform manner, regardless of the variety of modelling strategies found in real world resources. This homogenization of the problem space contribute to increasing the robusteness of the mediation systems with respect to the idiosyncrasies of the ontologies to be matched.
The use of linguistic resources from the Web and the discovery of suitable implementations of the provided interfaces causes that the applicability and perfomances of the ontology mediators increase as newer and better resources are publihsed. The success of MAPLE is therefore tied to the growth of the Linguistic Linked Open Data (LLOD) cloud, which collects the linguistic resources made available through the application of the Linked Data practices.
If you need to cite MAPLE in your scientific papers, journals and communications, please pick up the following paper.
- Manuel Fiorelli, Maria Teresa Pazienza and Armando Stellato A Meta-data Driven Platform for Semi-automatic Configuration of Ontology Mediators, In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC 2010) Reykjavik, Iceland, May, 26-31, 2014 bibtex