Identifying and Evaluating Ontology fragments for Mapping

This week, being the first, the focus was on identifying different ontologies related to Earth & Environmental Sciences. The emphasis of my internship is on evaluating and consequently implementing various algorithms present which can effectively map ontologies related to the said domain. The initial step in this direction is to first choose ontologies of importance and then identify the fragments which can be mapped using one or more algorithms.

The teleconference with the mentors really helped to put things in perspective. Various ontologies present in the stanford bioportal & the esip semantic portal were identified and my task was to do a sequential review of the same portals and possibly choose some more. As a result a good set is now in place. Various mapping techniques/algorithms are present and are listed in the IEEE paper on Ontology Matching (Pavel Shvaiko and Jerome Euzenat) and more in the Results of oaei12. I have been reading through these papers and trying to draw out relevant conclusions.

Points to ponder (as suggested by the mentors) while choosing an algorithm should revolve around scalability, and to see if the implementations are available as open source and basically to conclude that which algorithm shall serve in a particular situation. The first week has been exciting, to say the least.

Leave a Reply

Your email address will not be published. Required fields are marked *