Fuzzy Constraint-Based Schema Matching Formulation
Alsayed Algergawy, Eike Schallehn, Gunter Saake · 2001
Abstra t. The deep Web has many hallenges to be solved. Among them is s hema mat hing. In this paper, we build a on eptual onne tion between the s hema mat hing problem SMP and the fuzzy onstraint optimization problem FCOP. In parti ular, we propose the use of the fuzzy onstraint optimization problem as a framework to model and formalize the s hema mat hing problem. By formalizing the SMP as a FCOP, we gain many benets. First, we ould express it as a ombinatorial optimization problem with a set of soft onstraints whi h are able to ope with un ertainty in s hema mat hing. Se ond, the a tual algorithm solution be omes independent of the on rete graph model, allowing us to hange the model without ae ting the algorithm by introdu ing a new level of abstra tion. Moreover, we ould dis over omplex mat hes easily. Finally, we ould make a trade-o between s hema mat hing performan e aspe ts. Key words: s hema mat hing, onstraint programming, fuzzy onstraints, obje tive fun tion 1. Introdu tion. The deep Web (also known as Deepnet or the hidden Web) refers to the World Wide Web ontent that is not a part of the surfa e Web. It is estimated that the deep Web is several orders of magnitude larger than the surfa e Web [4℄. As the number of deep Web sour es has been in reasing as the eorts needed to enable users to explore and integrate these sour es be ome essential. As a result software systems have been developed to open the deep Web to users. S hema mat hing is the ore task of these systems. S hema mat hing is the task of identifying semanti orresponden es among elements of two or more