Possibilistic Networks for Aggregated Search in XML Documents

Abstract

In this paper, we are interested in aggregated search in conten-based XML information retrieval. Our goal is to revisit the granularity of the unit to be returned. More precisely, instead of returning the whole document or a list of disjoint elements of a document, we attempt to build the best element aggregation (set of non redundant elements) which is likely to be relevant to a query. For this, we present a model for XML information retrieval, based on possibilistic networks. The network structure provides a natural manner to represent the links between, a document, its elements and its content, and allows automatically selecting relevant and complementary elements. Experiments carried out on a sub-collection of INEX (INitiative for the Evaluation of XML retrieval), showed the effectiveness of the approach.