BUSCADORES SEMÁNTICOS PARA LA GESTION DEL CONOCIMIENTO

Abstract

The Semantic Web as it was designed by its creators could become the next revolution in Information society, similar to the WWW extension 20 years ago. This, so called, next generation Web was conceived for software instead of humans allowing human users delegating complex tasks to intelligent agents and applications. Making parallel analysis to the exploitation of the current Web this work identifies a set of challenges, mainly focusing on the challenge of the availability of a critical amount of semantically enabled content. This thesis introduces an acquisition system architecture for generating Semantic Web content. The goal of the described system is to process the current content of the WWW and upgrade it allowing for processing of software agents and applications. The state of the art introduces available technologies for information extraction as well as a set of similar working applications. The proposal starts with a hypothesis about the relations existing between the degree of structure present in the online sources and the technologies used for information extraction. In this sense, for instance, natural language processing becomes more efficient for texts where whole sentences or phrases are found, layout processing is really worthwhile for strongly visual structures (tables, lists, etc), and text engineering using all of them in a cooperation framework driven by some suitable strategy. The end of this memory includes a proposal for a Semantic Web application. It is a semantic portal using an implementation of the acquisition system for semantic content provision and allowing for advanced search and browse functionalities according to underlying model.