Large Scale Semantic Match ing: Agrovoc vs CABI

Achieving semantic interoperability is a difficult problem with a lot of challenges yet to address. Some of them include matching large-scale data sets, tackling the problem of missing background knowledge, evaluating large scale results, tuning the matching process and doing all of the above in a realistic setting with resource and time constraints. In this paper we report the results of a large-scale matching experiment performed on domain-specific resources: two agricultural thesauri. We share the experience concerning the above mentioned aspects of semantic matching, discuss the results, draw conclusions and outline perspective directions of future work.