Describes the PADOK-II project in Germany, which was designed to give information on the effects of linguistic algorithms on retrieval in a full-text database, the German Patent Information System (GPI). Relevance assessments are discussed, statistical evaluations are described, and searches are compared for the full-text section versus the…

Describes a project that was designed to test the relative efficacy of index terms and full-text for the retrieval of documents in MEDLINE journals. Precision and relevance of citations retrieved from full-text databases versus bibliographic databases is discussed, and the addition of MEDLINE indexing to enhance full-text medical journals is…

Describes and evaluates a learning model for information filtering and selective dissemination of information which is an adaptation of the generalized probabilistic model of information retrieval. The model is based on the concept of uncertainty sampling that allows for relevance feedback both on relevant and nonrelevant documents. (Author/LRW)

Increasing information retrieval performance using phrases and part-of-speech (POS) information is one example of a type of decision-making performance that is improved when using this linguistic information. The relative effectiveness of using multi-term phrases as opposed to individual terms is shown, as well as the relative worth of POS tagged…

Results from three studies examining 1295 relevance judgments by 36 information retrieval (IR) system end-users are reported. Both the region of the relevance judgments, from non-relevant to highly relevant, and motivations or levels for the relevance judgments are examined. Implications for relevance theory and IR systems evaluation are…

Discusses the need to develop the concept of user-based relevance for the benefit of users and for the meaningful development of future research in information retrieval. Characteristics of users' criteria of relevance are examined; and research methodology models are considered, including naturalistic inquiry versus scientific, or rationalistic,…

Reports on the performance of the vector space model in the presence of OCR (optical character recognition) errors in information retrieval. Highlights include precision and recall, a full-text test collection, smart vector representation, impact of weighting parameters, ranking variability, and the effect of relevance feedback. (Author/LRW)

Presents abstracts that discuss using data visualization for information retrieval and data mining, including immersive information space and spatial metaphors; spatial data using multi-dimensional matrices with maps; TREC (Text Retrieval Conference) experiments; users' information needs in cartographic information retrieval; and users' relevance…

Describes a scheme to represent a user's interest categories in an environment like the Web, and an adaptive algorithm to learn the dynamics of the user's interests through positive and negative relevance feedback. Proposes a three-descriptor model to represent a user's interests. Empirical studies confirm the effectiveness of this scheme to…

Reports the findings of an experiment using a simulated title page, author surnames, and title words, one-third of which were selected by each of the three authors, to determine the frequency of one-screen displays when used to search for known items in an implied Boolean retrieval system. (Author/AEF)

Investigates the information representation power of different modalities in the video data in order to collect empirical evidence for video surrogate creation and thus better support effective video browsing and information retrieval. Three types of video surrogates were created and compared under two user tasks--verbal comprehension and visual…

Describes a study that investigated the cognitive processes undergraduates used to select information while using a virtual library, GALILEO (Georgia Library Learning Online). Discusses higher order thinking processes, relevance judging, evaluation (critical thinking), decision making, reasoning involving documents, relevance-related reasoning,…

The Excerpta Medica on-line test data base of 181,653 records was found to be supplementary to MEDLINE and TOXLINE but especially responsive to searches on drug products. (Author)

A model of word occurrences in documents is presented in the context of a model information system. (Author/PF)

A statistical description of a text retrieval system allows the derivation of a unique and measurable performance criterion that, in turn, may be used in discussing various parameters of systems performance. (Author/PF)