English

Sequential change-point detection when unknown parameters are present in the pre-change distribution

Statistics Theory 2007-06-13 v2 Statistics Theory

Abstract

In the sequential change-point detection literature, most research specifies a required frequency of false alarms at a given pre-change distribution fθf_{\theta} and tries to minimize the detection delay for every possible post-change distribution gλg_{\lambda}. In this paper, motivated by a number of practical examples, we first consider the reverse question by specifying a required detection delay at a given post-change distribution and trying to minimize the frequency of false alarms for every possible pre-change distribution fθf_{\theta}. We present asymptotically optimal procedures for one-parameter exponential families. Next, we develop a general theory for change-point problems when both the pre-change distribution fθf_{\theta} and the post-change distribution gλg_{\lambda} involve unknown parameters. We also apply our approach to the special case of detecting shifts in the mean of independent normal observations.

Keywords

Cite

@article{arxiv.math/0605322,
  title  = {Sequential change-point detection when unknown parameters are present in the pre-change distribution},
  author = {Yajun Mei},
  journal= {arXiv preprint arXiv:math/0605322},
  year   = {2007}
}

Comments

Published at http://dx.doi.org/10.1214/009053605000000859 in the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)