At the Smithsonian Science Information Exchange, a computer system indexes word combinations in research summaries, according to a Classifying Dictionary, prior to review by the professional staff. (Author/PF)

A statistical method of drawing index terms or "highlights" from text is presented that uses no stop list, stem or other language- and domain-specific components, allowing operation in any domain or language with minimal modification. Some experimental results are presented, showing operation in six different languages. (67 references)…

A comparison was made of the performance in an automatic information retrieval environment of user queries and document abstracts available in natural language form in both English and French. The results obtained indicate that the automatic indexing and retrieval techniques actually used appear equally effective in handling the query and document…

This research reports an algorithmic approach to the automatic generation of thesauri for electronic community systems. The techniques used include term filtering, automatic indexing, and cluster analysis. The Worm Community System, used by molecular biologists studying the nematode worm C. elegans, was used as the testbed for this research.…

Describes automatic information retrieval system designed and built to handle Arabic data. Discusses cost-effectiveness of automatic indexing. Compares retrieval results using words as index terms versus stems and roots. Includes 19 tables; 60 queries using full words and relevance judgments are appended. (JAK)

The technique for machine-aided indexing developed at the Defense Documentation Center (DDC) is illustrated on a randomly chosen abstract. Additional text is provided in coded form so that the reader can more fully explore this technique. (2 references) (Author)

Describes an automatic method, based upon a text structure analysis and using a relatively small dictionary, to extract significant phrases from the title and the abstract of scientific and technical documents. The information representation of the document is discussed with relation to the construction of the document information retrieval…

The development of automatic indexing, abstracting, and extracting systems is investigated. Part I describes the development of tools for making syntactic and semantic distinctions of potential use in automatic indexing and extracting. One of these tools is a program for syntactic analysis (i.e., parsing) of English, the other is a dictionary of…

Grounded on object filtering, automatic indexing, and co-occurrence analysis, an experiment was performed using a parallel supercomputer to analyze over 400,000 abstracts in an INSPEC computer engineering collection. A user evaluation revealed that system-generated thesauri were better than the human-generated INSPEC subject thesaurus in concept…

Analyzes associations among different role operators quantitatively by taking input strings from 200 abstracts, each related to subject fields of taxation, genetic psychology, and Shakespearean drama, and subjecting them to the Chi-square test. Significant associations by other differencing operators and connectives are discussed. A schema of role…

Work has been performed on a fully automated approach for the general categorization of documents based on a training set of document titles and abstracts and journal-level indexing of the journals in which the documents are published. The goal of this research is to automatically index documents using a consistent, timely set of controlled…

Reports on the design and use of a computer system designed to assist online database searchers in determining the relevance of downloaded abstracts. The system, which runs on a personal computer, includes aids for individual review of abstracts and semiautomatic determination of relevance; prospects for fully automatic use are discussed. (CDD)

The first part of this dissertation is a metatheoretical discussion of the needs and means of semantic analysis. This discussion includes sections on metalanguage, deep and surface structure and structural semantics, and procedures for relating the English of abstracts to a classification concordance in the same language. The second part describes…

Short abstracts of works in progress or completed in the Department of Computing Science at the University of Alberta are presented under five major headings. The five categories are: Storage and search techniques for document data bases, Automatic classification, Study of indexing and classification languages through computer manipulation of data…

This paper argues that as the number of concordance lines presented to researchers increases in line with ever-growing corpus size, the automatic selection of the most central lines will become more important. A method will be presented for selecting the most representative members from a set of concordance lines on the basis of repeated lexical…