15 luglio 2009
In the Enterprise, extracting key entities such as names, locations, dates, and industry specific terminology from text can empower organizations to not only improve keyword search but also open the door to enhanced reasoning capabilities previously unavailable.
Using robust entity extraction technology allows for the automatic comprehension of words, sentences, paragraphs and whole documents. Categorization, extraction, domain establishment, taxonomy and ontology creation can then be built upon a Semantic Technology infra-structure to exploit the capabilities provided in OWL and RDF. This technology brings value to the Enterprise by allowing users to find, discover and create structured knowledge connections from what were previously unstructured information sets.
Expert System S.p.A. and Franz Inc. have paired their respective technologies to show how an advanced entity extractor and scalable RDF database can bring value to your project by unlocking previously unavailable knowledge from unstructured data.
The presentation will demonstrate how to represent output from the Cogito entity extractor in an AllegroGraph RDF database, and perform reasoning along with SPARQL queries. We will show an example of collecting 1000 articles scraping Google’s news page, and process them through Cogito to create RDF triples for use in AllegroGraph. The presentation will discuss tips and techniques in dealing with these representations, demonstrate how to relate entities to Linked Data such as DBpedia (or other external sources), and finally we will perform a number of queries on the resulting triple store data using some straight forward inferencing.