A data-intensive lightweight semantic wrapper approach to aid information integration.

Abstract

We argue for the flexible use of lightweight ontologies to aid information integration. Our proposed approach is grounded on the availability and exploitation of existing data sources in a networked environment such as the world wide web (instance data as it is commonly known in the description logic and ontology community). We have devised a mechanism using Semantic Web technologies that wraps each existing data source with semantic information, and we refer to this technique as SWEDER (Semantic Wrapping of Existing Data Sources with Embedded Rules). This technique provides representational homogeneity and a firm basis for information integration amongst these semantically enabled data sources. This technique also directly supports information integration though the use of context ontologies to align two or more semantically wrapped data sources and capture the rules that define these integrations. We have tested this proposed approach using a simple implementation in the domain of organisational and communication data and we speculate on the future directions for this lightweight approach to semantic enablement and contextual alignment of existing network-available data sources.