Access

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If You Use a Screen Reader

This content is available through Read Online (Free) program, which relies on page scans. Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.

ALMOST OPTIMAL SEQUENTIAL TESTS OF DISCRETE COMPOSITE HYPOTHESES

Georgios Fellouris and Alexander G. Tartakovsky
Statistica Sinica
Vol. 23, No. 4, Statistical Genetics, Change-point Problems, Random Fields and Sequential Analysis — a special issue in honor of Professor David Siegmund's 70th birthday (October 2013), pp. 1717-1741
Stable URL: http://www.jstor.org/stable/24310819
Page Count: 25
  • Read Online (Free)
  • Subscribe ($19.50)
  • Cite this Item
Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
ALMOST OPTIMAL SEQUENTIAL TESTS OF DISCRETE COMPOSITE HYPOTHESES
Preview not available

Abstract

We consider the problem of sequentially testing a simple null hypothesis, H0, versus a composite alternative hypothesis, H1, that consists of a finite set of densities. We study sequential tests that are based on thresholding of mixture-based likelihood ratio statistics and weighted generalized likelihood ratio statistics. It is shown that both sequential tests have several asymptotic optimality properties as error probabilities go to zero. First, for any weights, they minimize the expected sample size within a constant term under every scenario in H1 and at least to first order under H0. Second, for appropriate weights that are specified up to a prior distribution, they minimize a weighted expected sample size in H1 within an asymptotically negligible term. Third, for a particular prior distribution, they are almost minimax with respect to the expected Kullback–Leibler divergence until stopping. Furthermore, based on high-order asymptotic expansions for the operating characteristics, we propose prior distributions that lead to a robust behavior. Finally, based on asymptotic analysis as well as on simulation experiments, we argue that both tests have the same performance when they are designed with the same weights.

Page Thumbnails

  • Thumbnail: Page 
[1717]
    [1717]
  • Thumbnail: Page 
1718
    1718
  • Thumbnail: Page 
1719
    1719
  • Thumbnail: Page 
1720
    1720
  • Thumbnail: Page 
1721
    1721
  • Thumbnail: Page 
1722
    1722
  • Thumbnail: Page 
1723
    1723
  • Thumbnail: Page 
1724
    1724
  • Thumbnail: Page 
1725
    1725
  • Thumbnail: Page 
1726
    1726
  • Thumbnail: Page 
1727
    1727
  • Thumbnail: Page 
1728
    1728
  • Thumbnail: Page 
1729
    1729
  • Thumbnail: Page 
1730
    1730
  • Thumbnail: Page 
1731
    1731
  • Thumbnail: Page 
1732
    1732
  • Thumbnail: Page 
1733
    1733
  • Thumbnail: Page 
1734
    1734
  • Thumbnail: Page 
1735
    1735
  • Thumbnail: Page 
1736
    1736
  • Thumbnail: Page 
1737
    1737
  • Thumbnail: Page 
1738
    1738
  • Thumbnail: Page 
1739
    1739
  • Thumbnail: Page 
1740
    1740
  • Thumbnail: Page 
1741
    1741