Phénomènes financiers et mélange de lois : Une nouvelle méthode d’estimation des paramètres.

The main aim of this paper is to examine the qualities of the mixed diffusion-jump process whose parameters are random variables. The hypothesis of a Wiener geometric process applied to exchange rate has become doubtful at the beginning of the nineties, fact determined by a high leptokurtosis of the empirical distributions. The alternative of another distribution was studied in several articles. The mathematical model proposed in this paper has as fundamental hypothesis the fact that the distribution of the continuous part of the changes in the logarithms of exchange rate is a mixture of normals whose parameters are random variables.