Access by content of documents in an office information system

Abstract

This paper presents the integration of retrieval functions of an Information Retrieval System, IOTA, in an Office Information Server. Besides the linear scanning of the text (using a software and a hardware filter), two access methods are proposed. The first one is based on a simple indexing of documents based on signatures. Here, texts are treated as character strings. We call this method Textual Search. The second one is based on the extention of Signature Methods for implementing the Indexing Relation of IOTA, where meaningful terms (noun groups, for example) are identified in the text together with grammatical information. We call this method of signature computation the Indexing-Term Signature. The resulting access method is called Semantic Search. We present the current experimentations using the SCHUSS hardware filter as a scanning accelerator and the results of different alternatives of implementation of these Retrieval functions.