Eprep fact-file

What's the Problem?

One of the major problems on the Semantic Web is annotating gathered documents with semantically useful information.

The problem stems from the quantity of documents that require annotation and the reluctance of autohors to provide metadata in a machine readable way.

Towards a Solution

Eprep is a system for automatically adding metadata to Eprints submissions.

It works by accepting a URL that points to an HTML version of the paper and processing it with Gate. It then looks at the entites recognised by the Gate system and picks ones that are likly to be names of authours.

Process:

1. The user enters the URL in a field that appears on the Eprints page
2. Gate pulls out the Entity names
3. A script extracts the names that are likly to be authours
4. The Eprints page adds a dropdown box obove every authout metadata field on the submission page, containing all the names
5. When a name is selected some javascript in the page talks to the OKBC-PHP server and request and entities or type Person matching that name
6. The javascript then looks up the appropriate properties from the KB and fills in the textg fields

Screenshots

Semantic representation

View in the AKT Triplestore Browser or as RDF.

Also available in DOAP RDF (Description Of A Project)