Explains why Internet groups are a good medium for retrieving and exchanging information. An example of the use of netgroups to obtain information on the Gulf War of 1990-91 is given; the benefits of the informal nature of netgroups are discussed; and a mediated discussion group, Humanist, is described. (LRW)

Discusses the maximum entropy principle (MEP) and applies it to the design of probabilistic retrieval systems. Inconsistent expert judgments and the resulting optimization problem are examined, a linear programing problem is presented that checks the consistency of the expert judgments, and four schemes are proposed to find revised judgments. (12…

Discusses metasearch engines for information retrieval on the Internet, describes some of their advantages and disadvantages, and reviews the features of 10 popular metasearch engines currently available (Ask Jeeves, C4 Cyber 411, Dogpile, Highway 61, Inference Find, Mamma, Metacrawler, Metafind, Profusion, and Savvy Search). (LRW)

Shows that in advanced information-retrieval applications capabilities for data aggregation, transitive computation and non-first normal-form relational computation are often necessary at the same time. Topics include complex objects; advanced data models; query languages; query formulation; knowledge representation; and query-language syntax.…

Considers how ideas from document clustering can be used to improve retrieval accuracy of ranked lists in interactive systems and how to evaluate system effectiveness. Describes a TREC (Text Retrieval Conference) study that constructed and evaluated systems that present the user with ranked lists and a visualization of inter-document similarities.…

Presents an ongoing series of experiments with the TREC (Text Retrieval Conference) Interactive Track that test the feasibility and effectiveness of using clustering and classification as an aid to retrieval and answer construction. Results indicate that the success of the approach depends on assessing the quality of the final answers generated by…

Investigates the problem of fusing collections of documents provided by multiple information retrieval systems. A parameterized approach is suggested in which a parameter determines how the documents in the individual collections are interleaved to form a fused list of documents. (Author/LRW)

This session abstract discusses the history, philosophy, and theories of knowledge management to better understand its social and organizational potentials and limitations. Topics include determinacy of sense, information retrieval, and the Data Retrieval Model versus the Document Retrieval Model; discussions about knowledge; and surplus…

Explains vocabulary mining in information retrieval and describes a framework for vocabulary mining that allows the use of rough set-based approximations even when documents and queries are described using weighted, or fuzzy, representations. Examines coordination between multiple vocabulary views and applies the framework to the Unified Medical…

Discussion of word stemming for text processing and information retrieval focuses on an algorithm for generating stemmers from text stemmer specification files. Highlights include the use of finite state machines for stemming; stemmer specification file format; stemmer generation and driver algorithms; and stemmer timing comparisons. (Author/LRW)

Explains the SMART information retrieval project at Cornell University, which emphasizes completely automatic approaches to the understanding and retrieval of large quantities of text. Describes work in the sixth TREC (Text Retrieval Conference) that focused on routing, ad hoc, and foreign language environments; cross-lingual runs; and maintaining…

Discusses performance of the built-in search facilities of the commercially available Verity search engine at the sixth TREC (Text Retrieval Conference) and explores the impact of simple enhancements that could improve performance. Highlights include the ad hoc approach; Verity Boolean-based query language; term selection; term weighting;…

Explains how a search engine, or information retrieval system, operates. Four modules are described: the document processor, including term stemming, term weight assignment, and creating an index; the query processor, including parsing and query term weighting; the search and match function; and ranking capability. (LRW)

Considers information retrieval systems as a subset of information systems and examines the information spaces of systems and cognitive spaces of users. Presents a method for automatically generating information spaces from document collections that uses term cooccurrence, eigensystems analysis, and Principal Components Analysis, and discusses…

This article explores how Computer Sciences Corporation (CSC) manages knowledge retrieval by employees when they need to access documents written by colleagues in geographically distant units. CSC's establishment of virtual communities of practice facilitates the coordination of knowledge, and minimises contextual gaps between senders and…