|
|
Results and resourcesEuropeana Semantic LayerEuropeanaConnect significantly improves the possibility for Europeana users to discover the online collections of European museums, archives, libraries and audio-visual institutions by semantically enriching metadata. Thereby Europeana will be able to automatically suggest materials to the user that are related to a particular retrieved object. EuropeanaConnect also contributes to the Europeana Data Model (EDM). The following results and activities have been carried out so far:
Semantic LayerThe Semantic Layer built by EuropeanaConnect aims at providing a uniform, machine-actionable, web-enabled access to the reference knowledge capitalized by the various stakeholders of Europeana.eu. The objective is to make the metadata knowledge available for enabling the Europeana.eu users to benefit from semantics-intensive functions, as will be specified in other WP1 deliverables.
To enable semantic search a common data model was needed.
WP1 contributed to create the Europeana Data Model (EDM).
The EDM is aimed at being an integration medium for collecting,
connecting and enriching the descriptions/metadata provided by Europeana
content providers. As such, it includes any element (i.e., class or property)
found in a content provider's description/metadata. Giving an account of all
these elements is clearly an impossible task, since they form an open set, i.e.
a set that can be extended as new providers join the Europeana information space.
In the course of converting ESE (Europeana Semantic Elements) to the EDM, it was
decided to make this data available in the
Linked Open Data Cloud.
WP1 contributed in terms of technical and lobbying support.
Community Meetings to approve EDMv5 were held:
Semantic Search Engine for Europeana in the Thought lab pilot:The links between metadata records and structured vocabularies in the Semantic Layer can be exploited for semantic search. In semantic search, a machine uses these links to make associations and put search terms into context. For example, a search for 'Picasso' will lead to works created by Pablo Picasso as well as works depicting the painter. Artworks made of 'Picasso marble' are also retrieved and each of these contexts can be displayed in a separate cluster. Through the use of multilingual vocabularies, limited multilinguality is achieved. In this WIKI we track the vocabulary contributions: http://europeanalabs.eu/wiki/WP12Vocabularies The semantic search engine prototype is available in English, French and Dutch and contains data of the Rijksmuseum Amsterdam, the Musée du Louvre in Paris, and the Rijksbureau voor Kunsthistorische Documentatie (Netherlands Institute for Art History) in The Hague.
visit the Semantic Searching Prototype here: Semantic AlignmentsTo achieve the interconnectedness of Europeana resources, we need to establish links between the vocabularies that are used to annotate these resources. Rather than attempting to produce one unified ontology, our approach is to find alignments between the local vocabularies used to annotate the original data and more general pivot vocabularies.
AMALGAME Matching Tool
Fuzzy Alignment Tool
DSNotify Semantic Enrichment and Enrichment ExperimentsWhile for many Europeana metadata records terms from structured vocabularies are already used, a large number of metadata fields are filled with unstructured textual content. To further integrate these metadate records in the semantic layer, enrichment is needed to establish explicit semantic links between the representation of the collection objects (surrogates) and the (Europeana) vocabularies.In our approach, we employ the Amalgame tool to enrich these values by mapping them to existing vocabularies. In a first step, the values for a given metadata field are converted to a temporary vocabulary, which is then interactively aligned with the target vocabulary. The resulting correspondences are then stored in the semantic layer as metadata for a separate Europeana-owned aggregation, to preserve provenance data. RDF-based Metadata ManagementOrganizations often have difficulties to decide which solution they should adopt because comprehensive comparisons of existing RDF stores are hardly available and experiences w.r.t performance and scalability are still missing. This report summarizes the results of qualitative and quantitative study carried out on existing RDF stores in the context of the European Digital Library project.Europeana RDF Store Report
|