Statistical Science News

Dr Yvo Pokern: Nonparametric Drift Estimation for Stochastic Differential Equations

2 April 2012

In a paper recently accepted for publication in Biometrika, Papaspiliopoulos, Pokern, Roberts and Stuart have introduced nonparametric drift estimation for stochastic differential equations using novel partial differential equation methods. Applications to molecular dynamics and mathematical finance are also presented. The underlying probability theory requires new convergence theorems for the local time of diffusions on the circle and is presented in a separate paper (submitted). Possible extensions include second order hypoelliptic diffusions.