Detecting the presence of a random drift in Brownian motion, with motivating applications

Dr Peter Johnson (University of Manchester)

Date: Wednesday 11 December 3PM - Room 210, Mathematical Sciences Building

Abstract: Bayesian sequential testing and quickest detection problems for Brownian motion with a constant drift have been well studied. The presented work considers the case where the drift is random variable and only the distribution is known a priori.  We solve the sequential testing problem for a set of admissible laws which includes i) one-sided, and ii) two-point symmetric. This discussion will be motivated via applications in epidemiology and

bio-chemistry.

This talk follows the 2pm talk from William Turner (Imperial College) - Randomised path developments, and signature kernels as universal scaling limits

 

Back to: Institute for Financial and Actuarial Mathematics