Home        

People      

Projects    

Publications

Events      

Links       

Interactive Graphical Approach to Schema Matching (IGA)

With the increasingly emerging services provided through Web query interfaces across the Internet, interoperation among their services is becoming critically important. Finding semantic correspondences between elements of the schemas hidden behind these query interfaces is the crux to the interoperation. Automatic generation of schema mappings has been studied intensely and drawn much research interest recently in both XML and database research community. Little research however has been conducted in discovering and matching schemas behind these query interfaces. Taking the linguistic, data type and structure similarity into account, this paper proposes a new hybrid approach to the schema matching problem. The approach improves the matching accuracy by considering the semantic relationships of input schemas. The approach is also able to combine results of various matchers, and can be easily adapted to various domains. By accepting user customization, the approach provides an interactive environment for improving matching accuracy through the feedback. We have implemented a prototype and conducted several experiments to evaluate the matching algorithm. Results show that noteworthy improvements over simple schema matching have been achieved and the interactive environment provides friendly interface for users to tune the results.

IGA Demonstration

Please visit our online demo for IGA schema matching system.