You are here

Populating the Semantic Web with Historical Text

SSP Seminar Series
Kate Byrne

The Semantic Web, or Web of Data, is slowly growing and may be approaching the "critical mass" point at which its content will take off exponentially - much as the present "Document" Web did in the 1990s. Much of the data currently appearing in Semantic Web form, as RDF graph data, is new material generated specially; but of course there are vast amounts of older knowledge that could enrich the Web of Data.

This talk will look at aspects of my PhD work with cultural heritage data, which involves transforming existing archive content into an RDF graph dataset. Much of this content data (which is about Scotland's archaeological and architectural history) is in the form of free text documents, and the talk will concentrate on the process of trying to extract structure automatically from this. The approach is through named entity recognition followed by detection of text relations. The extracted relations can then be used to build an RDF graph that can be integrated with data from other sources.

Date and time: 
Tuesday, 9 December, 2008 - 11:00
60 minutes