A formal treatment of document information systems

摘要

This paper deals with a mathematical formulation of document information systems. The model of document information systems is given as an interpretation of a formal language. In the formal language, a concept of the labeled terms is introduced as the most basic concept, from a viewpoint of hierarchical semantic structure analysis of documents. The labels are intended to denote the context information, called the information functions in this paper, in the text of documents. By the introduction of the labeled terms, the expressive power of the description of semantic information of documents and queries has considerably been increased in comparison with a version of simple terms. Various conventional retrieval methods and more general retrieval method in the augmented framework are systematically described.