Reports on the development of an NLP (natural language processing) component for processing the free-text comments on life insurance applications for evaluation by an underwriting expert system. A sublanguage grammar approach with strong reliance on semantic word classes is described. Highlights include lexical analysis, adjacency analysis, and…

Introduces a special issue on research on hypertext systems and information retrieval. Systems design aspects, including navigation models, link weighting, multiple representation techniques, automatic authoring, user modelling, varying levels of abstraction, and evaluation techniques are addressed. (EA)

Discusses the automatic construction of abstracts from the texts of documents. Automatic sentence extraction is examined, including syntactic criteria and semantic relational criteria; textual cohesion is discussed, including the recognition of anaphors and noun phrases; text structure is addressed; and the concept of abstract-frames is described.…

Recent developments in cognitive science are assessed as challenge to well established view in philosophy of linguistics that meaning is inherent to language and is relatively fixed. Information provided by words and sentences, language comprehension, constructions of interpretations (not meanings), and arguments against relativism are discussed.…

Discusses and examines two approaches to concept-based retrieval of hypermedia information that improve a user's ability to navigate large systems--term indexing with semantically coupled thesauri and semantic hyperindexing. Visualization of index structures is discussed, and an example, the Cube of Contents retrieval model for the "Active…

Describes a method based on Bayesian networks for searching hypertext systems. Discussion covers the use of Bayesian networks for structuring index terms and representing user information needs; use of link semantics based on constrained spreading activation to find starting points for browsing; and evaluation of a prototype system. (64…

Discusses the interpretation of nominal compounds in the context of natural language processing for information retrieval. Knowledge-intensive algorithms that can successfully interpret compounds found in technical documents are described, and experiments are reported that indicate these algorithms may be unnecessary for improving retrieval…

Suggests structure based on two sets of principles for deriving meaning from data: Shannon's measure of entropy, which provides means of measuring amount of information in message; and Ranganathan's faceted classification scheme, which provides means of determining number of meaningful data. Syntax, meaning, and semantics of data are discussed.…