الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 132 |  |  |
| | | from | 132 | | |
Abstract
Due to the rapid growth of data sources, the need to data interoperability increases. And such interoperability can be achieved in a variety of ways such as Federated Database, data warehousing or peer Data Management Systems. All of these applications and specially Federated Database require schema integration to be able to provide the user with a unified view of the underlying data sources. However, schema integration process is complicated due to the heterogeneity between participating Database systems. The most challenged problem in schema integration is the identification of similar concepts that is called Schema matching. The most important challenges for schema matching is 1) providing a high degree of automation and minimize the manual work as possible , 2) identifies all possible matches (high quality results) , 3) provides a lower execution time . This thesis proposes a schema matching framework that consists of four phases dedicated to federated database system , the proposed framework aimed to minimize the manual work by only identifying the databases needed to be matched, it also aimed to provide a high quality results by exploiting all available data in the matching process and aimed to provides a better execution time and a lower resources consumption by providing an algorithm that adapts according to the complexity of the matching problem or the heterogeneity of the elements to be matched . The proposed framework has been tested on a simulation of federated database system to evaluate its effectiveness and efficiency. Experimental results consist of individual evaluation and comparative evolution, the individual experiments showed how the system adapts according to the matching complexity. The comparative evaluation showed that the system provides overall good results with lower execution time.