Accession Number : AD0285255

Title :   HIGH-SPEED DOCUMENT PERUSAL

Corporate Author : IBM WATSON RESEARCH CENTER YORKTOWN HEIGHTS N Y

Personal Author(s) : KOCHEN,M. ; ABRAHAM,C.T.

Report Date : 01 MAY 1962

Pagination or Media Count : 1

Abstract : This report includes: A SAMPLING PROCEDURE FOR CLUSTERING SIMILAR DOCUMENTS, by C. T. Abraham. Dec 6l. TIME ESTIMATION IN BOOLEAN INDEX SEARCHING, by E. Wong. Dec 61.AN ENGLISH-LIKE EXTENSION OF AN APPLIED PREDICATE CALCULUS, by H. Bohnert. Feb 62.AFOSR support led to the accomplishment of five results concerning the possibility of high-speed document perusal. These are: derivation of a formula for the average time to search an index; algorithms for translating English-like sentences into logic-like sentences; development of efficient techniques for grouping similar texts; implementation of a high-speed automatic dictionary lookup procedure; and, construction of computer programs for constructing representative abstracts and index terms. These constitute some first steps toward experimentally demonstrating the feasibility of a cooperative man-machine system for high-quality, high-speed perusal of large document collections; also, toward solving some of the basic logico-linguistic problems in the way of more completely automatic, highquality perusal. (Author)

Descriptors :   *DOCUMENTS, *INDEXES, *INFORMATION RETRIEVAL, *MATHEMATICAL COMPUTER DATA, *MATHEMATICAL LOGIC, AUTOMATIC, CODING, COMPUTERS, HUMAN FACTORS ENGINEERING, MACHINE TRANSLATION, MATHEMATICAL ANALYSIS, READING, READING MACHINES, SAMPLING, SEQUENCES(MATHEMATICS)

Distribution Statement : APPROVED FOR PUBLIC RELEASE