Next: Semantic Intensity Spectrum Up: CAPTURING REPRESENTING AND OPERATIONALISING Previous: CAPTURING REPRESENTING AND OPERATIONALISING

Survey

We did a thorough research in the literature and compiled a comprehensive survey on the state-of-the-art of semantic integration technologies. We point the reader to the 6th month report [1] for the detailed report published as a project deliverable. In this short section we simply recapitulate those findings in the form of summary tables with respect to some key issues: a classification regarding a stepwise process in applying semantic integration systems (table 1.1; the systems' objectives and capabilities (table 1.2; criteria for designing and developing semantic integration systems (table 1.3).

Classification of Semantic Integration Systems

In [1] we identified four phases of semantic integration:

(a) pre-integration preparation (a.k.a. normalisation)
(b) similarity discovery
(c) similarity representation (also includes reasoning)
(d) similarity execution (a.k.a. post-process)

The merit of such a representation is to provide the means for comparing different systems and technologies and putting them into context.

In table 1.1 we present such a comparison in a classification table. For instance, IF-Map focuses mainly on how to discover similarities between two ontologies and how to represent similarities, e.g. as RDF triples. On the other hand, FCA-Merge addresses only the similarity discovery.

Table 1.1: Classification of systems with respect to four phases of semantic integration.

Objectives of Semantic Integration Systems

A number of themes could be placed under the name semantic integration, ranging from federated database systems to distributed ontology development. Hence, it is beneficial to identify the objectives that a potential semantic integration is meant to serve.

In table 1.2 we summarise the major objectives for each system we reviewed in [1]. For instance, IF-Map is mainly for ontology mapping while CUPID is for database schema matching. Note that a system might have multiple major objectives.

Table 1.2: Objectives of semantic integration systems.

Criteria for developing Semantic Integration Systems

By carefully studying existing ontology mapping and database schema matching systems, we identified the following criteria that one could look into when developing a system for semantic integration: objectives, input, output, automation, extensibility, complexity and scalability.

We elaborate on each of these criteria in detail in [1] but here we summarise them in table 1.3.

Table 1.3: Semantic integration systems' features.

Bibliography

[1]: Y. Kalfoglou, B. Hu, D. Reynolds, and N. Shadbolt. Semantic Integration Technologies. 6th month deliverable, University of Southampton and HP Labs, ECS e-Prints report #10842, April 2005.

This material was prepared under the CROSI project. Copyright remains with the authors. Parts or the whole of this text have been published in conferences, workshops and other knowledge disseminating events.
CROSI presents this information online merely for sake of information dissemination.
This material should not be copy-pasted without acknowledging its origins.
Please contact the authors for information on how to use or reference this material.