الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 152 |  |  |
| | | from | 152 | | |
Abstract
Achieving semantic interoperability is a current challenge in the field of data integration in order to bridge semantic conflicts that occur when the participating sources and receivers use different or implicit data assumptions. Providing a framework that automatically detects and resolves semantic conflicts is considered a daunting task for many reasons. It should preserve the local autonomy of the integrated sources and provide a standard query language for accessing the integrated data on a global basis. The framework must also be adaptable- to accommodate changes gracefully, extensible -to allow easy addition and removal of data sources, and scalable- to handle large number of data sources with minor modifications.
Many existing traditional and ontology-based approaches have tried to achieve semantic interoperability. Traditional approaches have certain drawbacks that make them inappropriate for integrating information from a large number of data sources. Existing ontology-based approaches for semantic interoperability have not also been sufficiently effective because there is no systematic methodology to follow, no concert methodology for building ontologies and they use a static data model in reconciling the semantic
conflicts. Time-varying semantics can possibly be handled through view definitions that partition a data source into multiple ones. This is a manual process to be performed before query mediation, which increase the number of sources, making the conversions more difficult to maintain and affect the scalability of these approaches.
We propose semantic conflicts reconciliation (SCR) framework. It is an ontology-based system in which all data semantics are explicitly described in the knowledge representation phase and are automatically taken into account through the interpretation
mediation service phase so that conflicts are detected and resolved automatically at the query time. It provides a systematic methodology for describing assumptions through
attaching annotations to the merged ontology for each data element.