Data driven rank tests for classes of tail alternatives

Willem/Wim Albers, W.C.M. Kallenberg, F. Martini

Research output: Book/ReportReportOther research output

42 Downloads (Pure)

Abstract

Tail alternatives describe the frequent occurrence of a non-constant shift in the two-sample problem with a shift function increasing in the tail. The classes of shift functions can be built up using Legendre polynomials. It is important to rightly choose the number of polynomials involved. Here this choice is based on the data, using a modification of Schwarz's selection rule. Given the data driven choice of the model, appropriate rank tests are applied. Simulations show that the new data driven rank tests work very well. While other tests for detecting shift alternatives as Wilcoxon's test may completely break down for important classes of tail alternatives, the new tests have high and stable power. The new tests have also higher power than data driven rank tests for the unconstrained two-sample problem. Theoretical support is obtained by proving consistency of the new tests against very large classes of alternatives, including all common tail alternatives. A simple but accurate approximation of the null distribution makes application of the new tests easy.
Original languageUndefined
Place of PublicationEnschede
PublisherUniversity of Twente, Department of Applied Mathematics
Publication statusPublished - 1999

Publication series

Name
PublisherDepartment of Applied Mathematics, University of Twente
No.1504
ISSN (Print)0169-2690

Keywords

  • IR-65692
  • MSC-62E25
  • MSC-62G10
  • MSC-62G20
  • EWI-3324

Cite this

Albers, WW., Kallenberg, W. C. M., & Martini, F. (1999). Data driven rank tests for classes of tail alternatives. Enschede: University of Twente, Department of Applied Mathematics.
Albers, Willem/Wim ; Kallenberg, W.C.M. ; Martini, F. / Data driven rank tests for classes of tail alternatives. Enschede : University of Twente, Department of Applied Mathematics, 1999.
@book{39eb66296f81403f8bce498a359176ef,
title = "Data driven rank tests for classes of tail alternatives",
abstract = "Tail alternatives describe the frequent occurrence of a non-constant shift in the two-sample problem with a shift function increasing in the tail. The classes of shift functions can be built up using Legendre polynomials. It is important to rightly choose the number of polynomials involved. Here this choice is based on the data, using a modification of Schwarz's selection rule. Given the data driven choice of the model, appropriate rank tests are applied. Simulations show that the new data driven rank tests work very well. While other tests for detecting shift alternatives as Wilcoxon's test may completely break down for important classes of tail alternatives, the new tests have high and stable power. The new tests have also higher power than data driven rank tests for the unconstrained two-sample problem. Theoretical support is obtained by proving consistency of the new tests against very large classes of alternatives, including all common tail alternatives. A simple but accurate approximation of the null distribution makes application of the new tests easy.",
keywords = "IR-65692, MSC-62E25, MSC-62G10, MSC-62G20, EWI-3324",
author = "Willem/Wim Albers and W.C.M. Kallenberg and F. Martini",
note = "Imported from MEMORANDA",
year = "1999",
language = "Undefined",
publisher = "University of Twente, Department of Applied Mathematics",
number = "1504",

}

Albers, WW, Kallenberg, WCM & Martini, F 1999, Data driven rank tests for classes of tail alternatives. University of Twente, Department of Applied Mathematics, Enschede.

Data driven rank tests for classes of tail alternatives. / Albers, Willem/Wim; Kallenberg, W.C.M.; Martini, F.

Enschede : University of Twente, Department of Applied Mathematics, 1999.

Research output: Book/ReportReportOther research output

TY - BOOK

T1 - Data driven rank tests for classes of tail alternatives

AU - Albers, Willem/Wim

AU - Kallenberg, W.C.M.

AU - Martini, F.

N1 - Imported from MEMORANDA

PY - 1999

Y1 - 1999

N2 - Tail alternatives describe the frequent occurrence of a non-constant shift in the two-sample problem with a shift function increasing in the tail. The classes of shift functions can be built up using Legendre polynomials. It is important to rightly choose the number of polynomials involved. Here this choice is based on the data, using a modification of Schwarz's selection rule. Given the data driven choice of the model, appropriate rank tests are applied. Simulations show that the new data driven rank tests work very well. While other tests for detecting shift alternatives as Wilcoxon's test may completely break down for important classes of tail alternatives, the new tests have high and stable power. The new tests have also higher power than data driven rank tests for the unconstrained two-sample problem. Theoretical support is obtained by proving consistency of the new tests against very large classes of alternatives, including all common tail alternatives. A simple but accurate approximation of the null distribution makes application of the new tests easy.

AB - Tail alternatives describe the frequent occurrence of a non-constant shift in the two-sample problem with a shift function increasing in the tail. The classes of shift functions can be built up using Legendre polynomials. It is important to rightly choose the number of polynomials involved. Here this choice is based on the data, using a modification of Schwarz's selection rule. Given the data driven choice of the model, appropriate rank tests are applied. Simulations show that the new data driven rank tests work very well. While other tests for detecting shift alternatives as Wilcoxon's test may completely break down for important classes of tail alternatives, the new tests have high and stable power. The new tests have also higher power than data driven rank tests for the unconstrained two-sample problem. Theoretical support is obtained by proving consistency of the new tests against very large classes of alternatives, including all common tail alternatives. A simple but accurate approximation of the null distribution makes application of the new tests easy.

KW - IR-65692

KW - MSC-62E25

KW - MSC-62G10

KW - MSC-62G20

KW - EWI-3324

M3 - Report

BT - Data driven rank tests for classes of tail alternatives

PB - University of Twente, Department of Applied Mathematics

CY - Enschede

ER -

Albers WW, Kallenberg WCM, Martini F. Data driven rank tests for classes of tail alternatives. Enschede: University of Twente, Department of Applied Mathematics, 1999.