Full Text Searching and Information Overload

Abstract

This article classifies information retrieval applications into three classes depending on the correspondence between a user's request and the queries posed to the document base. It is argued that the mapping of requests (on a semantic level) to formalized queries (often on a lexical level) determines the range of retrieval effectiveness that may be obtained and that this classification may explain the discrepancy found in some information retrieval tests. It may also shed new light on a debate in the profession about the efficiency of retrieval systems in relation to precision, recall and information overload.