Abstract. The classification of products and services enables reliable and efficient electronic exchanges of product data across organizations. Many companies classify products (a) according to generic or industryspecific product classification standards, or (b) by using proprietary category systems. Such classification systems often contain thousands of product classes that are updated over time. This implies a large quantity of useful product category information for e-commerce applications on the Web of Data. Thus, instead of building up product ontologies from scratch, which is costly, tedious, error-prone, and high-maintenance, it is generally easier to derive them from existing classifications. In this paper, we (1) describe a generic, semi-automated method for deriving OWL ontologies from product classification standards and proprietary category systems. Moreover, we (2) show that our approach generates logically and semantically correct vocabularies, and (3) present the practical benefit of our approach. The resulting product ontologies are compatible with the GoodRelations vocabulary for e-commerce and with schema.org and can be used to enrich product and offer descriptions on the Semantic Web with granular product type information from existing data sources.

In this paper, we describe the technical approach of and experiences gained in exposing a major share of the European building and construction materials market on the basis of the GoodRelations Web vocabulary for e-commerce. This allows for the fine-grained search for products, suppliers, and warehouses for any building-related sourcing needs. Because building materials show a very high item specificity, they are very interesting for new types of search. Also, transportation costs for building materials are very significant, which makes the distance from the warehouse to the point of consumption a critical dimension of search. Based on existing data sources, we were able to include a rich, machinereadable description of individual product features using the FreeClassOWL ontology, which allow for multi-dimensional search. The result is one of the largest and richest public datasets for a well-defined trade sector that is available on the Semantic Web.

Abstract. The ongoing trend towards open data embraced by the Semantic Web has started to produce a large number of data sources. These data sources are published using RDF vocabularies, and it is possible to navigate throughout the data due to their graph topology. This paper presents LinksB2N, an algorithm for discovering information overlaps in RDF data repositories and performing data integration with no human intervention over data sets that partially share the same domain. LinksB2N identifies equivalent RDF resources from different data sets with several degrees of confidence. The algorithm relies on a novel approach that uses clustering techniques to analyze the distribution of unique objects that contain overlapping information in different data graphs. Our contribution is illustrated in the context of the Market Blended Insight project 1 by applying the LinksB2N algorithm to data sets in the order of hundreds of millions of RDF triples containing relevant information in the domain of business to business (B2B) marketing analysis.

There are ontology domain concepts that can be represented according to multiple alternative classification criteria. Current ontology modeling guidelines do not explicitly consider this aspect in the representation of such concepts. To assist with this issue, we examined a domain-specific simplified model for facet analysis used in Library Science. This model produces a Faceted Classification Scheme (FCS) which accounts for the multiple alternative classification criteria of the domain concept under scrutiny. A comparative analysis between a FCS and the Normalisation Ontology Design Pattern (ODP) indicates the existence of key similarities between the elements in the generic structure of both knowledge representation models. As a result, a mapping is identified that allows to transform a FCS into an OWL DL ontology applying the Normalisation ODP. Our contribution is illustrated with an existing FCS example in the domain of "Dishwashing Detergent" that benefits from the outcome of this study.

Abstract. Design-pattern driven ontology construction, whether manual or (partially) automated, relies on the availability of curated repositories of Ontology Design Patterns (ODPs) adequately characterized. In order to consistently apply a given ODP, not only it is important to characterize it in full, but also examine its alignment or deviation to other relevant ODPs in relation to it. Otherwise, possible inconsistencies in the application can lead to interoperability issues among the ontology models involved. In that context, this paper revisits a specific version of three different ODPs: Class as a Property Value (CPV), Value Partition (VP) and Normalisation. The review of the CPV identifies two distinct modelling problems being tangled that prompt to decouple the pattern into two variants: a strict and a coarse CPV pattern. The examination continues with a comparative analysis among the patterns that reveals key alignments and differences at the structural and semantic level. These findings extends the reusability and compositional characteristics of the strict and coarse variants of the CPV ODP in relation to the other two patterns. To illustrate our contribution existing examples in the literature are revisited. They demonstrate the alignments, differences and prototypical OWL idioms identified, which can assist ontology practitioners in mitigating the opportunity for inconsistencies when applying these recurrent ontology building blocks.