Meta-model for a large credit portofolio
Clément Rey  1@  
1 : Centre de Mathématiques Appliquées - Ecole Polytechnique  -  Website
Ecole Polytechnique, Centre National de la Recherche Scientifique : UMR7641
École Polytechnique Route de Saclay 91128 Palaiseau Cedex -  France

We expose a Meta-model for a large sum of indicator functions which depend on a common random factor. It is built from the chaos expansion with respect to this common random factor. We then propose a Gaussian approximation for the chaos coefficient which relies on a Central Limit Theorem established in this work. From a practical view point the Large sum of indicators can be seen as a Loss function in Cedrit Risk for which we propose a numerically efficient simulation using the Meta-model. 



  • Poster
Online user: 30 Privacy
Loading...