Supporting environmental scanning and organizational communication with the processing of text: the use of computer-generated abstracts

This research proposes a model text-based decision support system designed to support the activities of environmental scanning and organizational communication by actively filtering and condensing text. To filter textbased information requires the use of automatic routing schemes; to condense text requires the use of computergenerated abstracts or extracts. A key element in the model system is the ability of the computer to condense text by generating short abstracts of documents.

Two approaches to condensing text have been proposed: (1) using natural language processing techniques to construct a knowledge base of the document contents, from which to write an abstract, and (2) employing algorithm based extracting systems to generate extracts of important sentences and phrases. Systems using natural language techniques are still being researched; most are successful only in limited domains. Systems using extracting algorithms have been researched, but have not been applied to the problem of information overload in an organizational decision-making context. These two approaches were tested in a laboratory setting with student subjects.