Ranking invariance based on similarity measures in document retrieval

Abstract

To automatically retrieve documents or images from a database, retrieval systems use similarity measures to compare a request based on features extracted from the documents. As a result, documents are ordered in a list by decreasing correspondance to the request, Several comparison measures are used in the field and it is difficult to choose one or another. In this paper, we show that they can be grouped into classes of equivalent behavior, Then, in a query by example process, the choice of these measure can be reduced to the choice of a family of them. © Springer-Vorlag Berlin Heidelberg 2006.