A Survey of Cognitive Theories to Support Data Integration

Business intelligence applications are being increasingly used to facilitate managerial insight and maintain competitiveness. These applications rely on the availability of integrated data from multiple data sources, making database integration an increasingly important task. A central step in the process of data integration is schema matching, the identification of similar elements in the two databases. While a number of approaches have been proposed, the majority of schema matching techniques are based on ad-hoc heuristics, instead of an established theoretical foundation. The absence of a theoretical foundation makes it difficult to explain and improve schema matching process. This research surveys current cognitive theories of similarity and demonstrates their application to the problem of schema matching. Better integration techniques will benefit business intelligence applications and can thereby contribute to business value.