Describes a model for a front end query processor to deal with incomplete information in a database. The effectiveness of the processor in restricting the number of objects to be processed in a query and in aiding the interpretation of a query is discussed. (26 references) (Author/CLB)

Describes an algorithm for the calculation of term discrimination values that may be used when the interdocument similarity measure used is the cosine coefficient and when the document representations have been weighted using one particular term weighting scheme. (7 references) (Author/CLB)

Reports on a project showing that relational data models, including unnormalized models that allow the nesting of relations, cannot fully capture the structural properties of lexical information. A feature-based model that allows for a full representation of some nesting and defines a factoring mechanism is described and demonstrated. (38…

Presents a methodology for extending string indexing from index display production to the provision of descriptor phrases for term proximity searching. Objectives of the extension are discussed, and an illustration of the use of an existing string indexing system for this purpose is given. (14 references) (CLB)

Analyzes different kinds of reformulations used in information retrieval systems where full text databases are accessed through natural language queries. Tests of these reformulations on large full text databases managed by the Syntactic and Probabilistic Indexing and Retrieval of Information in Texts (SPIRIT) system are described, and an expert…

Describes SCISOR (System for Conceptual Information Summarization, Organization and Retrieval), a prototype intelligent information retrieval system that extracts useful information from large bodies of text. It overcomes limitations of linguistic coverage by applying a text processing strategy that is tolerant of unknown words and gaps in…

Explores the fundamentals of browsing behavior as a basis for the creation of an intelligent computer system to assist in online information retrieval. A prototype system is described, and the results of initial tests are reported. (5 references) (Author/CLB)

Presents a model for the a priori analysis of natural language queries which uses an algorithm to transform the query into a logical pattern that is used to determine the answerability of the query. The results of testing by a prototype system implemented in PROLOG are discussed. (20 references) (CLB)

Discussion of discourse modeling in artificial intelligence and computational linguistics focuses on a dialog model that incorporates the illocutionary aspects of information-seeking dialogs. The relevance of a theory developed for the analysis of written texts--Rhetorical Structure Theory--is explained, and future work is discussed. (25…

Describes the two basic approaches to the calculation of term discrimination values for automatic indexing. The results of an experiment that investigated the differences between algorithms of these two approaches in terms of their impact on the discrimination value model are reported and discussed. (13 references) (Author/CLB)

Describes experiments that compared a new method for automatically analyzing semantic structures in text by statistical means with the standard vector space model. Findings indicate that combining both methods improved performance over either alone. The effects of other experimental variables on retrieval performance (term weighting, suffix…

Describes the hyperterm system REALIST (Retrieval Aids by Linguistics and Statistics) with emphasis on its semantic component, which generates term relations from free-text input. Experiments with various similarity measures are discussed, and the quality of the associated terms is evaluated using term recall and term precision measures. (22…

Describes a computational model of the human process of generating text. The system design and generation process are discussed with particular attention to domain independence and cross language portability. The results of system tests are presented, the generator is evaluated with respect to current generators, and future directions are…

Argues that the best strategy for building integrated, personalized information systems consists of linguistically oriented knowledge and inference, rather than artificial intelligence techniques. The discussion addresses difficulties in the evaluation of such systems and strategies for devising performance studies. (five references) (CLB)

Proposes that topic representations should consist of excerpts from domain thesaurus, generated by spreading activation technique. Algorithm for generating excerpts is outlined and exemplified, and problem of assessing resemblance between two excerpts is discussed. Whether existing thesauri are adequate for this purpose is questioned, and ideas…