CMStatistics 2016: Start Registration
View Submission - CMStatistics
B1442
Title: Non-parametric Bayesian estimation of the diffusion coefficient Authors:  Shota Gugushvili - Leiden University (Netherlands) [presenting]
Abstract: A non-parametric Bayesian approach is considered to estimate the deterministic diffusion coefficient of a stochastic differential equation based on discrete time observations on its solution. On the theoretical side, we justify our approach by demonstrating that the posterior distribution asymptotically, as the sample size grows to infinity, concentrates around the `true' diffusion coefficient and derive the corresponding posterior contraction rate. On the implementational side, we show that our approach is straightforward to implement, requires little fine-tuning from the user, and leads to good practical results.