Study Of The Detecting And Resolving Data Conflict In Data Integration

Establishing semantic interoperability among heterogeneous information source has been a critical issue in the database community, however, the problem of interoperability depends on resolving data conflicts, this is the major task in data integration, data conflict includes Schema-level and Data-level conflict, and the latter is more difficult. In heterogeneous and distribute database, local database are independent of each other, each of them was managed by different administrator, so the data in different database are most likely have different semantic and different data value, In this case, It has result in the data source conflict in local databases.Eventually bring the wrong result of query.The problem of semantic interoperability has generally been tackled by one of two approaches:the federated schema approach and the domain ontology approach.The federated schema approach attempts to construct a federated schema and the participating local schemas.However,the drawback of this approach is that it was not designed to be indepengdent of particular schemas or applications.Thus,as new schemas or application join a community or as potential conflicts emerge,one may have to significantly modify the federated schema.On the other hand,the domain ontology approach uses a machine understandable definition of concepts and relationship between concepts so that there is a shared common understanding within a community.Its knowledge is domain specific,but indepengdent of particular schemas and applications.However,one needs additional tools actually capture and represent the knowledge needed to resolve semantic conflicts.Based on the work already existing, this paper studied the problem of detecting and resolving semantic conflicts both on schema-level and semantic-level. We proposed a more efficient and comprehensive approach to detect and resolve semantic conflicts based on the existing software environment, mean while a more comprehensive and rich semantics to meet the varieties of conflict representation and resolve. This approach includes two important parts:one ontology used to resolve schema-level data conflicts; the other ontology represents conflict classification used to resolve semantic-level data semantic conflicts. Currently, the system has been developed, experimental evaluations on which indicate that this approach can resolve the data conflicts effectively and keep independent of integration pattern and domain.The main contributions of this paper are:(1)Studying systematically the popular approaches about detecting and resolving semantic conflicts,Finding many shortcome of current work by detailed analysis andcompareision;(2)According the various requirement of practical data integration,we have improved the process of detecting and resolving semantic conflicts based on the exisiting model,partically improved transform algorithm between conflict concept,reduce the constrains of one-way transform in old approaches, meanwhile increased automatioin of whole process;(3)Based on the classification of semantic conflict,we have extended the semantic of database schema,so that our approach can detect and resolve more semantic conflicts;(4)Taking into account the convenience to deal with the problem,we defing two kind of ontologyes for schema-level and semantic-level data conflicts resprectively,so that we can improve the efficiency and also make sure resolving more data conflicts。...