Using repetition to increase precision in files with large blocks of text

Abstract

Searchers need to develop a set of heuristics (a variety of problem solving tactics and operations) for searching full‐text databases. Those heuristics which have evolved for searching bibliographic databases are often not effective when applied to full‐text. Presented here are methods which can be helpful in increasing precision when searching full‐text journal files for items on a given topic at BRS. Mead Data Central and Dialog. All of the techniques are based on achieving some repetition of central concepts as expressed in natural language. Research in automatic indexing and in full‐text retrieval provides a theoretical justification for a heuristic of this type. Schematic representations of strategies and sample searches which represent application of the techniques on each of the three systems are provided.

Citation

Jean McKinin, E., Ellen Sievert, M. and Diane Johnson, E. (1989), "Using repetition to increase precision in files with large blocks of text", Online Review, Vol. 13 No. 5, pp. 369-382. https://doi.org/10.1108/eb024317