Understanding The Schema Matching Problem

Schema matching plays the central role in many applications that require interoperability between heterogeneous data sources. The best way to attain comprehensive understanding of the schema matching problem is to construct a complete, if possible, problem formulation. Schema matching has been intensively researched and many matching systems have been developed. However, specifications of the schema matching problem being solved by these systems do not exist, or if it exists do not take uncertainty problems into account. In this paper, we propose the use of the fuzzy constraint problem (FCP) as a framework to model and understand the schema matching problem. In an effort to achieve more generic approach, we first transform the schema matching problem into a graph matching problem by transforming schemas to be matched into a common model namely rooted labeled graphs. Then, with the aid of this common model, we formulate the graph matching problem into a fuzzy constraint problem. By formalizing the schema matching problem as a FCP, we could express it as a combinatorial problem with soft constraints which enables us dealing with inherent uncertainty in schema matching. Key–Words: Schema matching, Rooted-labeled graphs, Constraint programming, Fuzzy constraints, Objective functions