Browse Books

Reviewer: Alessandro Berni

The use of semantic Web technologies for the extraction of knowledge from digital data is a topic of high interest. Its significance becomes evident when recalling the central role of knowledge as the basic currency of the information society. At the same time, the amount of digital information produced each year challenges current interpretation and preservation capabilities (a recent International Data Corporation (IDC) report estimates that the digital data produced in 2006, mainly in unstructured form, exceeded 160 Exabytes). By today's standards, it is common to share data across heterogeneous systems using Extensible Markup Language (XML), although significant human efforts are still required to synthesize the semantics that lay behind it, or, if you prefer, its meaning. Current improvements include the development of ontologies, composed of entities, relations, and axioms, that are capable of providing semantically rich representations of underlying data models. One unfortunate weakness of present approaches is that even ontologies developed using the same syntax (for example, using Web ontology language (OWL)) can represent the same concept with differences in naming or in structure. Ontology alignment therefore becomes a necessary step toward interoperability, to identify and exploit relations between different ontologies. That is where the excellent work of Marc Ehrig comes into play. In his book, he seeks to provide a concise methodology for aligning ontologies, improving process outcome by using encoded semantics and exploiting similarities between individual elements. The resulting procedure encompasses six major steps (feature engineering, search step selection, similarity assessment, aggregation, interpretation, and iteration) that can be supplemented by additional strategies matched to end-user applications. The book is organized into four major sections. The first part covers the theoretical aspects of ontology alignment, and current approaches for the alignment of well-structured entities, such as database schemas. A second part, representing the core of the contribution, introduces general processes for ontology alignment, and describes, in detail, the steps to be followed. Part 3 shows the application of ontology alignment within five real-life test cases, while Part 4 discusses the future of semantic alignment. A final section provides additional information on the ontologies used throughout the book, as well as statistical and graphical representations. Although the book has been written as a comprehensive work, different audiences (experts in the field or people willing to deepen their knowledge of the topic) can jump directly to sections of specific interest. In any case, readers should not neglect Part 2, where they will discover more about semantic alignment processes. Overall, this is an advanced-level book, providing valuable tools and guidance for practitioners and researchers. Online Computing Reviews Service