Schema Matchers ’ Performance Improvement

Abstract: The growing number of business solutions has lead to heterogeneity. This heterogeneity has several aspects one of which is the database nonuniformity. Several applications, such as the B2B e-commerce, require the efficient communication between these systems. The problem roots at the database representational ambiguities and can be resolved by the matching of database schemas. This task referred as schema matching, has become one of the main the focus points of business interest. Several possible solutions have been aired lately, though the scope includes some not yet resolved problems. This then have several detrimental effects on the performance. The solutions so far introduced need the human supervision as their result set is not always reliable. We have recognized that their performance is strongly dependent on their parameters and other factors. That is why we endeavored to find means of pre-run scenario optimization and propose methods which better harness the current solutions. According to our recommendation the schema matchers should be calibrated with a proper parameter set before their execution. The optimal choice of threshold is crucial, we have proposed a method which should handle the problem in a sophisticated way