English

Multivariate sequential analysis with linear boundaries

Statistics Theory 2007-06-13 v1 Statistics Theory

Abstract

Let {Sn=(Xn,Wn)}n0\{S_n=(X_n,W_n)\}_{n\ge0} be a random walk with XnRX_n\in \mathbb{R} and WnRmW_n\in \mathbb{R}^m. Let τ=τa=inf{n:Xn>a}\tau=\tau_a=\inf\{n:X_n>a\}. The main results presented are two term asymptotic expansions for the joint distribution of SτS_{\tau} and τ\tau and the marginal distribution of h(Sτ/a,τ/a)h(S_{\tau}/a,\tau/a) in the limit aa\to\infty. These results are used to study the distribution of tt-statistics in sequential experiments with sample size τ\tau, and to remove bias from confidence intervals based on Anscombe's theorem.

Keywords

Cite

@article{arxiv.math/0611678,
  title  = {Multivariate sequential analysis with linear boundaries},
  author = {Robert Keener},
  journal= {arXiv preprint arXiv:math/0611678},
  year   = {2007}
}

Comments

Published at http://dx.doi.org/10.1214/074921706000000608 in the IMS Lecture Notes--Monograph Series (http://www.imstat.org/publications/lecnotes.htm) by the Institute of Mathematical Statistics (http://www.imstat.org)