Page de Fabien Duchateau

FRBRpedia

The FRBR model has received much attention due to its potential for greatly improving user interaction with digital libraries. However, the amount of information found on the Web is far larger than in digital libraries. In this demo, we present an approach to transform web-based resources to a FRBR compatible form, a process known as FRBRization. The FRBRized collection is then linked to DBpedia, thus providing a basis for information sharing and verification.

Related publications

1. FRBR-ML: A FRBR-based Framework for Semantic Interoperability
   Journal of Semantic Web, 2012
   Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg and Maja Zumer
   
   
   @article {semweb12,
     author = {Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg and Maja Zumer},
     title = {FRBR-ML: A FRBR-based Framework for Semantic Interoperability},
     journal = {Journal of Semantic Web},
     publisher = {IOS Press},
     year = {2012},
     volume = {3},
     number = {1},
     month = {jan},
     url = {http://dx.doi.org/10.3233/SW-2012-0044},
     issn = {1570-0844},
     pages = {23-43},
   }
   
   
2. FRBRPedia: a Tool for FRBRizing Web Products and Linking FRBR Entities to DBpedia
   Joint Conference on Digital Libraries (JCDL), 2011
   Fabien Duchateau and Naimdjon Takhirov and Trond Aalberg
   
   
   @inproceedings {jcdl11-demo,
     author = {Fabien Duchateau and Naimdjon Takhirov and Trond Aalberg},
     title = {FRBRPedia: a Tool for FRBRizing Web Products and Linking FRBR Entities to DBpedia},
     booktitle = {Joint Conference on Digital Libraries (JCDL)},
     year = {2011},
     publisher = {ACM},
     pages = {455-456},
     ee = {http://doi.acm.org/10.1145/1998076.1998183},
   }
   
   
3. Supporting FRBRization of Web Product Descriptions
   Theory and Practice of Digital Libraries (TPDL), 2011
   Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg
   
   
   @inproceedings {tpdl11-frbr,
     author = {Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg},
     title = {Supporting FRBRization of Web Product Descriptions},
     booktitle = {Theory and Practice of Digital Libraries (TPDL)},
     publisher = {Springer},
     year = {2011},
     pages = {69-76},
     isbn = {978-3-642-24468-1},
   }
   
   
4. Linking FRBR Entities to LOD through Semantic Matching
   Theory and Practice of Digital Libraries (TPDL), 2011
   Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg
   
   
   @inproceedings {tpdl11-lod,
     author = {Naimdjon Takhirov and Fabien Duchateau and Trond Aalberg},
     title = {Linking FRBR Entities to LOD through Semantic Matching},
     booktitle = {Theory and Practice of Digital Libraries (TPDL)},
     publisher = {Springer},
     year = {2011},
     pages = {284-295},
     isbn = {978-3-642-24468-1},
   }
   
   

Prototype

FRBRpedia is a GreaseMonkey plugin for web browsers. It enables to extract and convert products from Amazon bookstore into the FRBR model. We link the generated FRBR works to Linked Open Data cloud (LOD) by using entity matching techniques. One of the goals with linking to LOD is to verify the FRBRized information. We have used the DBpedia knowledge base since it contains potentially a great amount of information.

Installation steps for the plugin:

1. Install GreaseMonkey
2. Download frbr2amazon plugin
3. Browse Amazon product display page (e.g. the Da Vinci Code)

Details of the process:

• Based on the product information extracted from Amazon bookstore, we first need to create (at least) four FRBR entities: work, manifestation(s), expression(s), and actor(s).
• The first step consists of identifying the work. We use the external service OCLC Classify API to fulfill this goal. We send a query (e.g., ISBN number) to retrieve the work corresponding to the product. Although OCLC includes a large collection of works, it occurs that the service does not return any result, for instance with products specific to a country. In the future, we plan to integrate other methods for identifying a work such as z39.50.
• The second step of our FRBRization process deals with the discovery of related manifestations for the previously identified work. Similarly, we chose an OCLC Service, xISBN, since a search in the same database increases our chances of finding related manifestations. The xISBN Web service returns ISBNs and other information associated with an individual intellectual work that is represented in the WorldCat catalog.
• From the set of related manifestations, we can automatically generate the third category of FRBR entities: expressions. By analyzing attributes such as language and translator, we are able to create expressions, whose identifier is automatically generated.
• Next, the actors who contributed to this work are identified through the Virtual International Authority File (VIAF), a joint project between national libraries. This service aims at gathering authoritative names from many libraries into a global service freely available on the Web.
• Finally, the work, its expressions, manifestations and actors are linked during the last step. The final output of our FRBRization process is a set of RDF files for each entity type.

Screenshots