Journal of the American Society for Information Science · 1981 · 30 citations · 12 references
ComputingComputer ScienceDatabasesInformation RetrievalKnowledge RepresentationBoolean AlgebraBoolean RetrievalCriteria WeightingDocument RepresentationEvaluation Of Retrieval SystemsFuzzy LogicBoolean Document RepresentationBoolean ExpressionsBoolean Retrieval ModelProbabilistic Retrieval ModelsQuery EvaluationSearch Result RankingWeighting Systems
Abstract A major deficiency of traditional Boolean systems is their inability to represent the varying degrees to which a document may be written on a subject. In this article we isolate a number of criteria that should be met by any Boolean system generalized to have a weighting capability. It is proven that only one weighting rule satisfies these conditions—that associated with fuzzy‐set theory—and that this weighting scheme satisfies most of the other properties associated with Boolean algebra as well. Probabilistic weighting is then introduced as an alternative approach and the two systems compared. In the limit of zero/one weights, all systems considered converge to traditional Boolean retrieval.
12
L. A. Zadeh · Information and Control · 1965 · 66.1K citations
On Relevance, Probabilistic Indexing and Information Retrieval
M. E. Maron, J. L. Kuhns · Journal of the ACM · 1960 · 901 citations · Full text
Introduction to Probability Theory.
H. C. Fryer, Paul G. Hoel, Sidney C. Port et al. · Journal of the American Statistical Association · 1972 · 808 citations