Circular Likelihood Ratio Tests for the Main Problem

  • Johan C. Akkerboom
Part of the Lecture Notes in Statistics book series (LNS, volume 62)

Abstract

In the present chapter we discuss some ways in which the theory of Ch. 3 can be used to obtain simple and satisfactory procedures for testing problems with a polyhedral-cone-shaped alternative. While dealing with the main problem \( \left( {{H_0},{{\tilde H}_1}} \right) \) with r>2, in which the alternative is determined by the pointed polyhedral cone K, one may proceed as follows. First the circular cone C = C(a, w) is specified as a substitute for K, next a particular test is selected for the modified problem \( \left( {{H_0},{{\tilde H}_1}} \right) \) of testing against C\(0 r), and finally this test is performed as a solution to the original problem (H0 , H1). Assuming that the choice of test for (H0 , \( {\tilde H_1} \)) is made in advance, the gist is to choose the axis a and the axial angle w of the substitute C for K, cf. (3.0.1). The special case r=2, where K itself is a circular cone, can be considered a benchmark for the comparison of various approaches to the main problem.

Keywords

Covariance Resid Hull Eter Gout 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer-Verlag Berlin Heidelberg 1990

Authors and Affiliations

  • Johan C. Akkerboom
    • 1
  1. 1.Department of Statistical MethodsCentral Bureau of StatisticsHeerlenThe Netherlands

Personalised recommendations