Discusses alternatives to the Boolean system that would enable users of online bibliographic databases--primarily those containing scientific and technical information--to obtain their required level of performance at lowest cost. Highlights include recall and precision factors (natural and controlled language, titles, abstracts, quorum function)…

Presents decision-theoretic models which intrinsically include retrieval of multiple documents whereby system responds to request by presenting documents to patron in sequence, gathering feedback, and using information to modify future retrievals. Document independence model, set retrieval model, sequential retrieval model, learning model,…

Evaluates overlapping coverage of 21 scientific databases used in 10 online pesticide searches in an attempt to identify minimum number of databases needed to generate 90 percent of unique, relevant citations for given search. Comparison of searches combined under given pesticide usage (herbicide, fungicide, insecticide) is discussed. Nine…

Discusses information retrieval techniques involving both Boolean retrieval schemes and a weighting mechanism to produce ranked search output and points out theoretical ambiguities and inconsistencies demonstrated by recent research. An alternative search approach is outlined and analyzed and a more rigorous methodology is suggested. Eleven…

Six uses of classification in online retrieval systems are envisioned--to improve recall and precision; to conceptualize meaning of vague search terms; to provide structure for browsing; to provide framework for retrieval of nonbibliographic information; to collocate citations; and to achieve compatibility of retrieval languages. Seven sources are…

Presents information retrieval design approach in which queries of computer-based system consist of sets of terms, either unweighted or weighted with subjective term precision estimates, and retrieval outputs ranked by probability of usefulness estimated by "maximum entropy principle." Boolean and weighted request systems are discussed.…

The low rate of relevant titles (fewer than 50 percent) retrieved by online searching of keywords in titles in study conducted at the Mechanized Information Center, Ohio State University, is attributed to three sources: titles themselves, user ignorance of subject vocabulary, and general language problems. Eight references are cited. (EJS)

Formulates a model to accommodate the hypothesis that, if the content representation items of documents in a database are neither totally independent nor highly correlated, combining them in a certain way may increase information retrieval effectiveness. The results of a small experiment that tested this hypothesis are presented. (Author/JD)

Describes an experiment in passage retrieval for scientists using a variety of forms of retrieval question and search words selected independently by two different people. (FM)

The number of hits on Internet keyword searches can be limited by following some tips and tricks to using search engines efficiently. Recommends being specific, checking relevancy, using multiple search engines, and using directories. Includes a bibliography of search engines, directories, and print and Internet articles. (PEN)

Describes the use of three alternative methods for ranking films for information retrieval (IR). A large film-person incidence matrix is generated using the principle cast, directors, producers and screenwriters for each film. These attributes are used to measure film-film distances by creating a distance matrix. Distance, product distance, and…

Reports preliminary results of a study to develop and automate new metrics for assessment of information quality in text documents, particularly in news. Through focus group studies, quality judgment experiments, and textual feature extraction and analysis, nine quality aspects were generated and applied in human assessments. Experiments were…

Textual properties of documents were analyzed to identify predictive variables for various document qualities by means of statistical and linguistic methods. A collection of 1000 documents was created, and each document was judged in terms of nine document qualities (accuracy, reliability, objectivity, depth, author/producer credibility,…

Discusses ranking techniques for document retrieval and proposes an evaluation, or filtering, technique that reduces costs and memory usage by early recognition of which documents are likely to be highly ranked. Highlights include index design, ranked query evaluation, reducing the number of accumulators, inverted file structures, and experimental…

Discussion of information retrieval and relevance focuses on mutual information, a measure which represents the relation between two words. A model of a natural-language information-retrieval system that is based on a two-level document-ranking method using mutual information is presented, and a Korean encyclopedia test collection is explained.…