A probabilistic model for interactive retrieval is presented. Bayesian statistical decision theory principles are applied: use of prior and sample information about the relationship of document descriptions to query relevance; maximization of expected value of a utility function, to the problem of optimally restructuring search strategies in an…

Swets' theory of information retrieval allows the threads of document weighting formulae, probabilistic measures of effectiveness, and management theory, to be woven into a coherent pattern. Benefits are: the beginnings of a quantitative description of retrieval languages; destinction between retrieval systems and language; and question generality…

The model assumes that the relevance assigned to a document by a juror is a random variable. (PF)

Examines the main mathematical approaches to information retrieval, including both algebraic and probabilistic models, and describes difficulties which impede formalization of information retrieval processes. A number of developments are covered where new theoretical understandings have directly led to improved retrieval techniques and operations.…

Introduces a new model of viewing documents based on citing-cited relationship between them. Using Bayes' decision theory, it is shown how source document may be indexed and weighted by relevant cited document features, corresponding to one pass relevance feedback Model 1 (probabilistic indexing) or Model 2 (probabilistic retrieval). (24…

Presents adaptation of a probabilistic theoretical model previously used in relevance feedback for use in automatic indexing of documents (in the sense of imitating) human indexers. Methods for model application are proposed, independence assumptions used in the model are interpreted, and the probability of a dependence model is discussed.…

This report explains the following mathematical techniques which may be used for relating search requests to documents contained in a library: (1) Boolean Algebraic Retrieval, (2) Linear Statistical Retrieval, (3) Statistical Association Techniques for expanding a query and/or for expanding the set of retrieval documents, (4) Vector Space…

Presents and discusses the results of research into similarity measures for search request formulations which employ Boolean combinations of index terms. The use of a weighting mechanism to indicate the importance of attributes in a search formulation is described. A 16-item reference list is included. (JL)

Retrieval experiments with the Cranfield collection of 1,400 documents used strategies based on a probabilistic model for an initial search and intermediate search when no relevant documents were known. Results show initial strategy is better than conventional strategies in retrieval effectiveness and number of queries needed to retrieve…

Discusses the interdependencies between parameter estimation and properties of probabilistic models, such as dependency assumptions, binary vs. nonbinary features, and estimation sample selection. An optimum estimation for binary features applicable to information retrieval is defined, a method for computing this estimation using empirical data is…

Reports on a test of the applicability of the theory of General Scalability to documents in terms of transitivity, substitutability, and independence. Implications for the construction of test collections in information retrieval research are examined. (24 references) (EAM)

Describes a theoretical model based on binary vectors that was developed to improve relevance in information retrieval systems. Earlier probabilistic models are examined, quadratic and linear discriminant functions are discussed, relationships between queries and documents are considered, and further research is suggested. (15 references) (LRW)

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,…

Five models which perform search tasks of the Data Base Access and Search Environment (DBASE) are presented and used for comparison with subject search performance--unconstrained (rule-based, decision theoretic) and constrained (optimal rule-based, rule-based, probabilistic rule-based). Previous experiments and the modeling approach are briefly…