Etude et mise en œuvre des aspects itératifs de l’extraction de règles d’association dans une base de données

Abstract

In this paper, we propose a general framework for iterative computation of mining queries based on condensed representations. To this end, we first adapt the standard notions of maximal, closed and key patterns introduced in previous works, to the more general case of sets of mining queries defined by monotonic and anti-monotonic selection predicates. Then, we show in a particular case how condensed representations of a set Q of mining queries can be used : (i) to optimize the computation of the answer of a new mining query q, (ii) to efficiently modify the condensed representation of the answer of Q to obtain a condensed representation of the answer of Q ∪ {q}. Iterative Computation of Mining Queries based on Condensed Representations