Title: Pretty predictable models
Authors: Tommi Sottinen - University of Vaasa (Finland) [presenting]
Abstract: A class of stochastic processes that are generated by a so-called invertible Gaussian Volterra process is considered. By this we mean that we can, in an adaptive way, recover an underlying Brownian motion that generates the same filtration as the stochastic process under consideration. By using the underlying Brownian motion, we construct explicitly the regular future conditional law of our stochastic process conditioned on the past. Examples include fractional Brownian motions and geometric fractional Brownian motions.