| The elaboration of semantic matching between heterogeneous data sources is a fundamental step in the design of data sharing applications. This task is tedious and often error prone if handled manually. Therefore, many systems have been developed for its automation. However, the majority of them target a specific type of applications and rely on rigid approaches applying, without distinction, the same matching technique with the same fixed procedures on data to be aligned without regard for their characteristics. The problem is that such mapping systems are not adapted to the diversity of Internet where data sources and domains are numerous, highly heterogeneous and often changing. Moreover, virtually all solutions currently focus on the problem of finding simple (one-to-one) matching. This is likely due to the fact that complex (many-to-many) matching raises a far more difficult problem since the search space of concept combinations can be tremendously large.;Key words: context analysis, complex semantic matching, adaptive semantic matching, multi-strategy.;To better cope with data source heterogeneity, we propose INDIGO, a system which can compute simple and complex matching by taking into account data sources' context. The distinctive feature of INDIGO is to enrich data sources with semantic information extracted from their individual development artifacts. Thanks to this enrichment step, INDIGO can then compute a more accurate mapping between the two data sources thus enhanced. To generate the mapping, INDIGO relies on an adaptive multi-strategy matching approach which takes into account the diversified nature of data sources. It provides multiple matching strategies which can be flexibly combined to take into consideration the specificities of the data sources being aligned. |
