Download 05-2

WS Kendall, JM Marin and CP Robert

Brownian confidence bands on Monte Carlo output

Date: April 2005

Abstract: When considering a Monte Carlo estimation procedure, the path produced by successive partial estimates is often used as a guide for informal convergence diagnostics. However the confidence region associated with that path cannot be derived simplistically from the confidence interval for the estimate itself. An asymptotically correct approach can be based on the Brownian motion approximation of the path, but no exact formula for the corresponding area-minimizing confidence region is yet known. We construct proxy regions based on local time arguments and consider numerical approximations. These are then available for a more incisive assessment of the Monte Carlo procedure and thence of the estimate itself.

Keywords: boundary crossing probability; Brownian motion; central limit theorem; local time; Monte Carlo path; prediction region; simultaneous confidence region.