Merging Ontologies using a Bottom-up Lexical and Structural Approach

Information integration is the focus of different research domains for several years. With the emergence of the Internet as a very large database, this topic became of extreme relevance for a larger audience, seeking for mechanisms to align information from different data sources according to semantic needs and constraints. The classical integration approaches do not satisfy new operational requirements, thus new strategies should be proposed and developed. We suggest the adoption of a light alignment mechanism without merging data source. The alignment process occurs at ontological level, setting the components and transformation functions necessary to translate data from source to the target. Establishing the alignment process at ontological level allows the system to reason about the semantic embodied in the information and thus enhance the alignment results, according to both the source and target specifications. The resulting specifications are then applied in the transformation procedure, executed independently of data source entities.