INFORMATION STORAGE AND RETRIEVAL

Abstract : Some experimental procedures for the identification of information content, using word frequency counting techniques are described. An attempt is made in particular to determine those areas in a natural language text which contain more than an average amount of new information. The use of tree structures for the representation of relations between terms included in classification systems, and between words in the natural language is discussed. Procedures are also suggested for the automatic identification of structural relations, and for the use of trees to perform the matching process. Some problems connected with the use of syntactic analysis for the identification of document content, and various strategies which appear useful for the processing of structured information are discussed. Methods are described for the efficient representation of tree structures in computer storage. Programs are also exhibited to perform a variety of information retrieval operations. (Author)