PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Indice Probabiliste Discriminant de vraisemblance du lien pour des Données Volumineuses
Israel-César Lerman and Jérome Azé
RNTI Volume 1, Number Numéro spécial RNTI-E-1 "Mesures de qualité pour la fouille de données", pp. 69-94, 2004.

Abstract

The likelihood of the link probabilistic index, measuring an association rule, becomes no finely discriminant when the data size becomes enough large. The aim of this paper consists in showing the discriminant extension of this probabilistic index in order to measure an association rule in the context of a set of association rules. This method has been proposed for a long time and has been extensively validated in the framework of the AVL (Analyse de la Vraisemblance des Liens) hierachical clustering method of descriptive attributes. An experimental design is considered in order to establish the relevancy of our statistical approach. This latter is also theoritically validated.

EPrint Type:Article
Additional Information:Association rule, probabilistic discriminant index, validation
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Learning/Statistics & Optimisation
ID Code:649
Deposited By:Jérome Azé
Deposited On:29 December 2004