A Knowledge-Based Framework for Data Integration

Aiming at integrating heterogeneous data sources into a logic unified data view, data integration has drawn considerable attention. This paper proposes a novel knowledge-based framework for data integration within which schema integration and content integration are two focal modules. A constructive semantic recommendation algorithm is put forward for schema integration, and a rule-based cleansing procedure is summarized for content integration. Experiments validate the effectiveness and efficiency of the proposed algorithms and thus verify the framework is feasible and pragmatic.