Accession Number:

AD0724977

Title:

Scientific and Technical Information. Series 2. Information Processes and Systems. Number 6, 1969 (Selected Portions),

Descriptive Note:

Corporate Author:

FOREIGN TECHNOLOGY DIV WRIGHT-PATTERSON AFB OHIO

Personal Author(s):

Report Date:

1971-02-17

Pagination or Media Count:

33.0

Abstract:

The methods and results of a study to develop an IR language for automatic systems handling polytechnical documents are described. The descriptor dictionary includes both general and special terms expressed by words and phrases which contributes to better recall and precision figures it comprises a classified and a lexico-semantic index as well as generic and specific relations tables. The dictionary size is 5,542 descriptors and 3,073 keywords. The document indexing procedure includes the following steps document content analysis and description by natural words forming of the search pattern by using the descriptor dictionary. The techniques are described which are applied to the analysis of a document from different conceptual aspects constituting the elements of a formalized model of its condensed content. Conversion into IR language involves the use of the words both from the title and the text of a document. The essentials of the technique used for evaluation of the retrieval efficiency by applying statistical methods are set forth. Tests on a multi-subject collection revealed the possibility of a system operating at 85 percent recall and 7 percent relevance, with a standard deviation of 25 percent. Author

Subject Categories:

  • Information Science

Distribution Statement:

APPROVED FOR PUBLIC RELEASE