Thèse de doctorat en Informatique
Sous la direction de Gilbert Saporta.
Soutenue en 2006
à Paris, CNAM .
Pas de résumé disponible.
Combined use of association rules mining and clustering methods
The aim of this thesis is to reveal new information on quality by applying advanced statistical methops to a large database. Tis dataset contains vehicle defects detected during the manufacturing process. In the course of our sudy, we have emplyed the analogy of the shpping basket to analyse data using association rules mining. We use this metod to search for enormous and impractical number of rules. As a result, validating our findings becomes an impossible task. To solve this problem we propose to combine the use of clustering methods and association rules mining. This approach allows us to significantly reduce both the number and complexity of rules obtains. These rules can then be evaluated statistically using relevant indexes. To accompany this study, we present a graphical tool to help compare index properties. Our results have been confirmed by experts in the field