Building and interpreting term dependencies using association rules extracted from Galois Lattices
Abstract
This article deals with the text mining activity applied to technological watch. We focus on the extraction process of association rules using a Galois lattice and relate our experimentation on a corpora. The corpus is composed of abstracts of scientific texts extracted from a documentary database. The interpretation of the results is based upon an analysis of the different levels of the lattice as well as an analysis of the different interpretations associations rules can be assigned to.