Using background contextual knowledge for documents representation

DOI10.1007/3-540-63620-X_59MaRDI QIDQ4376801zbMATH OpenFDO

Publication date 11 February 1998

Published in Lecture Notes in Computer Science (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/3-540-63620-X_59

contextual dependencies feature selection techniques machine learning numerical taxonomy

Information storage and retrieval of data (68P20) Knowledge representation (68T30) Theory of languages and software systems (knowledge-based systems, expert systems, etc.) for artificial intelligence (68T35) Computing methodologies for text processing; mathematical typography (68U15) Parallel algorithms in computer science (68W10)

Review text

The authors follow two main arguments within the paper: (1) Making use of contextual dependencies between words in document representation captures important information necessary to work on document processing problems. (2) Whilst feature selection techniques from numerical taxonomy and machine learning produced good results in text categorization, the background knowledge is proved to be able to replace and/or supplement these techniques. These facts are shown in document categorization trials on the Reuters-22173 standard document collection. The obtained results indicate performances comparable to the best known results for the considered test set. For the entire collection see [Zbl 0879.00058].

Recommendations

Cited in

(2)

This page was built for publication: Using background contextual knowledge for documents representation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4376801)