Week two revolved around identifying and surveying the algorithms and systems available for matching of the ontologies. The focus was on studying the several systems that have appeared and repeatedly participated (in recent years) in the OAEI campaigns. The Ontology Matching paper (Shvaiko & Euzenat) discusses some of the systems which also have archival publications and hence a complete account of these works is available.
I concentrated my efforts on these matching systems/algorithms. All of these can deal with OWL, with many capable of handling RDFS or SKOS. Few are equipped with native GUI while some rely on the SEALS platform (being actively promoted by OAEI). The paper gives a brief description of Falcon, SAMBO, DSSim, RiMOM, ASMOV, Anchor-Flood and AgreementMaker. I went ahead and read the actual publications of these systems (at-least the introduction and intended purpose).
After an initial evaluation and discussions with the Mentors it transpires that efforts now should be directed towards exploring the Falcon and AgreementMaker systems. Reasons being the availability of proper documentation and GUI for these. Falcon deals with large ontologies and is open source, but is hosted over the SEALS platform. It can partition ontologies for effective matching and also can discover alignments. Its efficiency has been verified (see page 11 of the ontology matching paper). AgreementMaker, on the other hand, consists of a wide range of ontology and schema matchers(syntactic & structural). It has a neat GUI and it fairs well on important evaluation measures (precision, recall).
I intend to work on these two systems and maybe one another and conclude as to which of these suits best to the ontologies of our immediate interest. Also attached with this post are some slides which give a brief review of the above mentioned Ontology Matching Systems. Please have a look and any advice or recommendation on the use of these systems would be appreciated.