Operations Research Applied to Document Indexing and Retrieval Decisions

Abstract

This paper begins with a review of earher work in which a model of word occurrence formed the basis of a decision-making procedure for indexing or, more generally, retrieving documents in response to a request In the earlier work words were considered individually This paper extends the earher model to include interactions among terms The elaborated model allows one to decide whether to retrieve a document by taking into consideration occurrences of all the words in the text Retrieval in response to Boolean expresstons IS also considered, as are procedures for ranking documents in accordance with their assessed relevance to a request The discussion is within the framework of Bayesian decision theory