Deployment of the text-mining tool Bilbo on Revues.org
Bilbo identifies the bibliographic references in journal articles and semantisizes their constituent parts. It then identifies the DOIs corresponding to these references and, where they exist, adds them to the end of the reference as a hyperlink, making it possible to directly access the cited resource. Developed by OpenEdition Lab, Bilbo is now deployed on almost 80% of the journals on the Revues.org platform.
How does Bilbo work?
Bilbo (“Bibliographical Robot”) is a piece of software that detects, identifies, analyses and encodes bibliographic references in articles. Bilbo uses data mining and machine learning to identify the first name and surname of the author(s) and the title, publisher, year and place of publication of each bibliographic reference. The first version of Bilbo focuses on article bibliographies. A second version will extend the identification of bibliographic references to footnotes. Finally, a third stage will involve identifying implicit references within the body text of articles.
Bilbo will regularly analyse the same bibliographies as the algorithm develops, but also in light of the fact that DOIs are attributed to thousands of new publications each day.
Automatically identified references
Using the author and title of an article, Bilbo can query the search engine maintained by Crossref, the official registry for Digital Object Identifiers (DOIs), whose database contains millions of academic references. Bilbo can thereby retrieve the article’s DOI, where it exists, and add it to the reference in the article’s bibliography. The DOI is added as a hyperlink, allowing the reader to directly access the cited resource.
Once the reference is identified, Bilbo is able to enrich it with complementary data and display it in different formats. Readers consulting an article via a library or institution that has subscribed to one of OpenEdition’s Freemium programmes will be able to download references for which Bilbo has found DOIs in APA, MLA and Chicago formats. The list of subscribing libraries and institutions can be consulted on this page: <http://cleo.openedition.org/pilotage/abonnes>
Who runs Bilbo?
Developed by OpenEdition Lab, Bilbo is a research and development programme launched in 2011 that aims to develop features related to reading, writing, navigation and system recommendations. Two teams work closely together on the project: the OpenEdition team and the Sciences and Systems Laboratory team (LSIS, Aix-Marseille University – CNRS). Initial funding for the project was provided by a Google Grant for Digital Humanities.
- Recommendations and information about Bilbo for editorial teams on La Maison des Revues (in French): <http://maisondesrevues.org/680>
- The OpenEdition Lab research blog, which traces the progress of the Bilbo project since its inception (in French): <http://lab.hypotheses.org/category/bilbo-bibliographical-robot>
- A description of how DOIs work, on La Maison des Revues (in French): <http://maisondesrevues.org/253>.