Extraction of facet terms from article titles and their display in tabular form

Abstract

A trial was carned out to effect automatic extraction of significant words from titles of scientific papers and display them in tabular form. The extraction algonthm is based on the simple syntactic patterns developed around a "pivot word." which has high control over the use of functional words in its vicinity. A pivot word and prepositions as well as punctuation marks delimit and extract title words; no word dictionary is used. The approach was applied to article titles in analytical chemistry. which include "determination(s)" and "determining" as pivot words. The syntactic patterns were expressed by Backus Naur Form notations, from which extraction rules were denved. Sample titles were processed on a personal computer, loaded with LISP programs for information extraction and table format display in accordance with the facets essential to analy sis. The adequacy of the approach was regarded as good with regard to its simplicity. This approach not only extracts keywords from article titles but also provides them with proper roles in conformity with the facets pertinent to the subject concerned.