INFORUM 2006 - 12. Annual Conference On Professional Information Resources

Fulltext Information Extraction

Author: Jan Žbirka,  Tovek Partner, Czech Republic

Fulltexts

Abstract

The information extraction can be one of the important tools in the fulltext processing. The input text (frequently output of the fulltext search) is prepared by any preprocessor for the morpho-syntactic analysis. The main extraction process in the next step is the replacement of certain tokens, or words with a concept name. These concept names can be referenced in the new concepts at a higher level. The selected parts of the original text are consecutively replaced by concept name structures. What is simple and what is problematic, how to integrate this technology to fulltext analytic system and in which areas we can use it? It is discussed in this paper.

About the author

Graduated at the Charles University, Prague (library and information science). Worked as a programmer at the Prague Airport and at the Information Centre for Chemistry. Assistant professor at the University of Economics, Prague.

Was a member of the National Information Centre council. Was a member in the team localising fulltext technology Verity Inc. Founder of an electronic archive of the Economia publishing. Analyst as a Tovek Partner (fulltext and visualisation technologies).