Automating schema matching is challenging. Previous approaches (e.g. [MBR01, DDH01]) to automating schema matching focus on computing direct element matches between two schemas. Schemas, however, rarely match directly. Thus, to complete the task of schema matching, we must also compute indirect element matches. In this paper, we present a framework for generating direct as well as many indirect element matches between a target schema and a source schema. Recognizing expected data values associated with schema elements and applying schema-structure heuristics are the key ideas to computing indirect matches. Experiments we have conducted over several real-world application domains show encouraging results. Keyword: Schema matching, data integration, schema integration, data exchange.
Li Xu, David W. Embley