Describes the System for Conceptual Information Summarization, Organization, and Retrieval (SCISOR), which is designed to parse and understand short newspaper stories on corporate takeovers. Highlights include issues involved in developing memory organization and retrieval components, theoretical considerations involved in the system design, and a…

Introduces the concept of cover coefficient to indexing theory and discusses a way of computing term discrimination values by using the cover coefficient concept, index-term selection, and weight modification. This method is compared to an approach that uses similarity coefficients. (Author/EM)

Describes GRANT, an expert system for finding sources of funding given research proposals. The architecture of GRANT and the implementation of constrained spreading activation (a modified search algorithm based on semantic memory) are described, and recall and precision rates are analyzed. (Author/LRW)

Discusses expert systems and their value as components of information retrieval systems related to semantic inference, and describes IOTA, a model of an intelligent information retrieval system which emphasizes natural language query processing. Experimental results are discussed and current and future developments are highlighted. (Author/LRW)

A comparative evaluation of two computerized reference retrieval systems, INSTRUCT (a statistically-based system) and PLEXUS (a knowledge-based system) examined their retrieval effectiveness for a set of 19 queries and 512 documents on gardening. Best results were obtained by using terms suggested by PLEXUS as the basis for an INSTRUCT search. (36…

Describes a project that is developing an intelligent search intermediary to help users locate relevant passages in large full text databases. (The system automatically reformulates Boolean queries to improve search results and presents retrieved passages in decreasing order of estimated relevance.) The current status of the project and plans for…

Summarizes the similarities between classification systems for catalog selection and information retrieval systems, and discusses system characteristics that allow the use of measures such as recall and precision in system evaluation. (27 references) (Author/CLB)

Discusses general issues of computer and human understanding; contrasts three paradigms of information retrieval methodology, including statistical, deep semantic or natural language, and smart Boolean; describes CONIT, a knowledge-based intermediary retrieval assistance system; and examines system evaluation procedures, including a…

Describes a document retrieval system that provides for user interaction at several search stages to acquire a detailed specification of the user's information need. The use of domain knowledge to refine the query and a browsing mechanism are described in detail, and further research questions are identified. (CLB)

Presents a method of modeling the relevance relationship in information retrieval to answer the question of the theoretical limits of certain statistical methods. Hypergeometric probability distribution is used to construct an abstract model of a database of MEDLINE records, and results of tests of vector retrieval methods are reported. (28…

Study results indicate that it is feasible to develop automatic name recognition algorithms to distinguish character strings representing names from other character strings occurring in English language titles. This finding offers cautious promise for alleviating some of the labor intensive work of cataloging. (16 references) (Author/SD)

Provides a method for estimating precision without examining individual database documents, which requires knowledge of only the query or expressed information need and a set of database parameters constant for all queries. Concepts of historic and expected precision are examined, and applications of precision-document graphs are assessed. (36…

Discusses intelligent filtering systems and describes work in the INFOS (Intelligent News Filtering Organizational System) project designed to automatically categorize data as relevant or irrelevant based on user interests. The filtering is performed by combining keyword-based hill climbing, knowledge-based conceptual representation via WordNet,…

E-Referencer is an expert intermediary system for searching library online public access catalogs (OPACs) on the World Wide Web. It is implemented as a proxy server that mediates the interaction between the user and Boolean OPACs. It transforms a Boolean OPAC into a retrieval system with many of the search capabilities of Web search engines.…

Describes the use of PROLOG to program knowledge-based information retrieval systems, in which the knowledge contained in a document is translated into machine processable logic. Several examples of the resulting search process, and the program rules supporting the process, are given. (10 references) (CLB)