Machine Extracting Program (Russian).
Abstract
Development and implementation of a system for automatic extracting of technical Russian are described. The system is based in part upon the previously developed CLUE technique for machine extracting of English and Russian text. This technical report contains a discussion of the present system in comparison to previous methods, a description fof the operating system, and illustrations of machine, target, and random extracts used in a system evaluation. Experiments which were performed to develop some of the system components are described. A report of the results of a linguistic investigation to determine the eefficacy of certain linguistic properties of the text as aids to automatic extracting is included. The operating system incorporates two new extracting techniques. The first of these, the Cue Finder, is a self-teaching computer routine designed to select the optimum set of cue words from a set of documents. The second is a location model for selection of extract sentences in terms of their locations in the paragraph, the section indicated by a subheading, and the entire document. Also included are conclusions based on the study results and recommendations for subsequent development of the system. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1969
- Accession Number
- AD0853186
Entities
People
- Clyde E. Miller
- George M. Motherwell
- Richard K. Overton