What happens in Turin stays in Turin? We have a FREME report

FREME_Stoitsis

Freme-blog

Have you ever used services like Open Calais ? If yes then probably you were amazed by how easy is to automatically extract semantic information directly from the submitted text. If you are not familiar with Open Calais it is a service that extracts events, people, organisations, topics and social tags from unstructured data. I would suggest to try it. A lot of developers and companies are using such services to enrich their content, to improve discovery services, to build recommendations services and to extract analytics from content.

But could such great service work with any open dataset that you have for your domain and not only with a specific and close data? And could such service enrich the extracted information with information from linked data sources like DBpedia? This could be a truly democratic approach for extracting semantic information from open data since anyone could submit the data of his preference and could take advantage of other existing generic and domain specific open datasets.

This is one of the main goals that the FREME project wants to achieve. To build such an open approach for extracting semantic information form open data. The FREME aims at providing a platform that will make available in a friendly way such services.

AK participates in the FREME project as a company that will use the results to enable the multilingual sharing and discovery of agricultural research information. Using the FREME e-services we aim at enhancing our sharing and data reuse services of the AKStem product.

 

More specifically we  are using the FREME e-services to enrich the metadata records for the agricultural bibliographic information that we are collecting from data providers. Just to give you a specific example: we are receiving metadata records in which the name of a specific author may have several variations e.g. Giannis Stoitsis, John Stoitsis or G.S. Stoitsis. This results in a bad user experience as the specific filter will show the three different values for the same person. It is also not possible to browse through the publications of a specific author. Using the FREME e-services we can achieve two things:

  1. To disambiguate the author by comparing his name against an authority file like the one maintained by ORCID and
  2. To enrich the metadata record by introducing also ORCID uri for the author based on which we can get more info about the specific person and thus we can show this info to the end user of our services

We do similar enrichments for organisations using the Virtual International Authority File (VIAF) and places using the Geopolitical ontology. We can also add missing terms from AGROVOC that could help in multilingual browsing and searching of the bibliographic resources.

We participated in the second plenary meeting of the FREME project that took place in Turin on Thursday 3/9 and Friday 4/9 at the very nice premises of ISMB. We discussed the progress and first results from applying FREME e-services in our workflow. Stay tuned for updates on our work.

Leave a Comment.

Can you prove that you are not a bot? * Time limit is exhausted. Please reload CAPTCHA.