Abstract
. In this article we treat about classification of a set of data described by probability distributions. This classification can be obtained by mixture decomposition, the constituent ones of this one being obtained using parametric copulas. The difficulty of implementation of this method lies in the calculation of the densities related to the copulas, and this for dimensions higher than 2. We provide a simple response using the Clayton's copula..
Original language | French |
---|---|
Title of host publication | XXXV èmes Journées de Statistiques |
Place of Publication | Lyon |
Publisher | Société Française de Statistique |
Pages | 377-380 |
Number of pages | 4 |
Publication status | Published - 2003 |