A context-based approach for linguistic matching

As currently implemented by most data mapping systems, linguistic matching often boils down to string comparison or to a synonym look-up in a dictionary. But these solutions have proved to be inefficient for dealing with highly heterogeneous data sources. To cope with data source heterogeneity more efficiently, we introduce INDIGO, a system which computes semantic matching by taking into account data sources' context. The distinctive feature of INDIGO consists in enriching data sources with semantic information extracted from their individual development artifacts before mapping them. As explained in this article, experiments conducted on two case studies proved the relevance of this approach.