ON THE CONDITIONAL DISTRIBUTIONS AND THE EFFICIENT SIMULATIONS OF EXPONENTIAL INTEGRALS OF GAUSSIAN RANDOM FIELDS

• 成果类型: Article
• 署名作者: Liu, Jingchen; Xu, Gongjun
• 署名单位: Columbia University; University of Minnesota System; University of Minnesota Twin Cities
• 刊物名称: ANNALS OF APPLIED PROBABILITY
• ISSN/ISSBN: 1050-5164
• DOI: 10.1214/13-AAP960
• 发表日期: 2014
• 页码: 1691-1738
• 关键词: time-series; monte-carlo; maximum; regression; crossings; models; sums

摘要

In this paper, we consider the extreme behavior of a Gaussian random field f(t) living on a compact set T. In particular, we are interested in tail events associated with the integral integral(T) e(f(t)) dt. We construct a (non-Gaussian) random field whose distribution can be explicitly stated. This field approximates the conditional Gaussian random field f (given that integral(T) e(f(t)) dt exceeds a large value) in total variation. Based on this approximation, we show that the tail event of integral(T) e(f(t)) dt is asymptotically equivalent to the tail event of sup(T) gamma(t) where gamma(t) is a Gaussian process and it is an affine function of f(t) and its derivative field. In addition to the asymptotic description of the conditional field, we construct an efficient Monte Carlo estimator that runs in polynomial time of log b to compute the probability P (integral(T) e(f(t)) dt > b) with a prescribed relative accuracy.