Semantic techniques for improving keyword-based searching
- Autores: Raquel Trillo Lado
- Directores de la Tesis: Eduardo Mena Nieto (dir. tes.) , Nieves R. Brisaboa (tut. tes.)
- Lectura: En la Universidade da Coruña ( España ) en 2012
- Idioma: inglés
- Tribunal Calificador de la Tesis: Isidro Ramos Salavert (presid.) , Antonio Fariña (secret.) , Arantza Illarramendi Echave (voc.) , Asunción Gómez Pérez (voc.) , Francesco Guerra (voc.)
- Texto completo no disponible (Saber más ...)
- Resumen
In the last decade, Keyword-based Searching systems have become very popular mainly due to the success of web search engines such as Google or Yahoo!. Generally, these systems are based on the use of syntactic techniques that ignore the meaning of words (their semantics) and on the construction of specialized indexes of the content directly accessible. Due to these facts, the high semantic heterogeneity makes it difficult for users to locate the available information, and there is a huge amount of information (usually stored in structured sources) that is hidden to searching systems. Therefore, many researchers of several areas (Semantic Web, Deep Web, Databases, etc.) study these topics from different perspectives.
In this work, we tackle the application of semantic techniques to keywordbased search scenarios. Firstly, we study semantic measurements and their application to annotate and classify the results provided by a traditional search engine. Then, we focus on techniques to translate keyword queries into queries written in a formal unambiguous language in order to query structured data sources, in particular data sources based on Description Logics and relational data sources. We also present the prototypes Doctopush, QueryGen, and Keymantic, and experimental results that support the feasibility of our proposals.
