Matstat seminarium fredagen  23 februari, 2001

Monte Carlo density estimates for diffusion processes


Roger Pettersson, Växjö

Abstract:

The density at a final time point of a diffusion
process is approximated by Euler simulations. Instead of using just the
final values, we save also the other values of the paths to get a finer
density estimate. The idea is to use an integration by parts formula for
Malliavin derivatives. Comparison with traditional kernel density
estimates is made. 

Joint work with A Kohatsu-Higa, UPF, Barcelona.