16 Pages Posted: 17 Aug 2004
Date Written: August 2004
Changes of semantics in data sources further complicate the semantic heterogeneity problem. We identify four types of semantic heterogeneities related to changing semantics and present a solution based on an extension to the Context Interchange (COIN) framework. Changing semantics is represented as multi-valued contextual attributes in a shared ontology; however, only a single value is valid over a certain time interval. A mediator, implemented in abductive constraint logic programming, processes the semantics by solving temporal constraints for single-valued time intervals and automatically applying conversions to resolve semantic differences over these intervals. We also discuss the scalability of the approach and its applicability to the Semantic Web.
Keywords: semantic heterogeneity problem, Context Interchange (COIN) framework
Suggested Citation: Suggested Citation
Zhu, Hongwei (Harry) and Madnick, Stuart and Siegel, Michael, Representation and Reasoning about Changing Semantics in Heterogeneous Data Sources (August 2004). MIT Sloan Working Paper No. 4497-04; CISL Working Paper No. 2004-14. Available at SSRN: https://ssrn.com/abstract=577242 or http://dx.doi.org/10.2139/ssrn.577242