Advertisement

Computational Statistics

, Volume 16, Issue 4, pp 519–538 | Cite as

The split-up algorithm: a fast symbolic method for computing p-values of distribution-free statistics

  • Mark van de WielEmail author
Article

Summary

Many distribution-free statistics have the drawback that computing exact p-values under the null hypothesis is an intensive task. When the sample sizes are small or the number of ties is large, approximations are often unsatisfactory. Moreover, tables of exact critical values are not available for conditional rank statistics (ties, censoring), for rank statistics with arbitrary regression constants, or for permutation test statistics. In those cases, it is important to have a fast algorithm for computing exact p-values. We present a new algorithm and apply it to a large class of distribution-free one-sample, two-sample and serial statistics. The algorithm is based on splitting the probability generating function of the test statistic into two parts. We compare the speed of this “split-up algorithm” to that of existing procedures and we conclude that our new algorithm is faster in many cases.

Keywords

Exact p-values two-sample rank tests signed rank tests rank correlation tests rank autocorrelation tests permutation tests generating functions 

Notes

Acknowledgements

I would like to thank Alessandro Di Bucchianico and the referees of this journal for their useful comments on earlier versions of this paper.

References

  1. Gibbons, J.D., and J.W. Pratt (1975). P-values: interpretation and methodology. American Statistician, 29, 20–25.zbMATHGoogle Scholar
  2. Good, P.I. (1994). Permutation Tests, a Practical Guide to Resampling Methods for Testing Hypotheses. Springer-Verlag, New York.zbMATHGoogle Scholar
  3. Hallin, M., and G. Mélard (1988). Rank-based tests for randomness against first-order serial dependence. Journal of the American Statistical Association, 83, 1117–1128.MathSciNetCrossRefGoogle Scholar
  4. Hallin, M., and M.L. Puri (1988). Optimal rank-based procedures for time series analysis: testing an ARMA model against other ARMA models. Annals of Statistics, 16, 402–432.MathSciNetCrossRefGoogle Scholar
  5. Hallin, M., and M.L. Puri (1991). Time series analysis via rank order theory: signed-rank tests for ARMA models. Journal of Multivariate Analysis, 39, 1–29.MathSciNetCrossRefGoogle Scholar
  6. Kendall, M.G., and A. Stuart (1977). The Advanced Theory of Statistics, Volume 2. Charles Griffin & Co., London.zbMATHGoogle Scholar
  7. Mehta, C.R., N.R. Patel, and L.J. Wei (1988). Constructing exact significance tests with restricted randomization rules. Biometrika, 75, 295–302.CrossRefGoogle Scholar
  8. Pagano, M., and D. Tritchler (1983). On obtaining permutation distributions in polynomial time. Journal of the American Statistical Association, 78, 435–441.MathSciNetCrossRefGoogle Scholar
  9. Prentice, R.L., and P. Marek (1979). A qualitative discrepancy between censored data rank tests. Biometrics, 35, 861–867.CrossRefGoogle Scholar
  10. Puri, M.L., and P.K. Sen (1985). Nonparametric Methods in General Linear Models. Wiley, New York.zbMATHGoogle Scholar
  11. Streitberg, B., and J. Röhmel (1986). Exact distributions for permutation and rank tests: an introduction to some recently published algorithms. Statistical Software Newsletter, 12, 10–17.Google Scholar
  12. Streitberg, B., and J. Röhmel (1987). Exakte Verteilungen für Rang-und Randomisierungstests im allgemeinen c-Stichprobenproblem. EDV in Medizin und Biologie, 18, 12–19.Google Scholar
  13. Van de Wiel, M.A. (1996). Computing exact distributions of rank statistics with computer algebra. Master’s thesis, Eindhoven University of Technology, The Netherlands.Google Scholar
  14. Van de Wiel, M.A. (1998). Exact distributions of two-sample rank statistics and block rank statistics using computer algebra. Technical Report Memorandum COSOR 98-14, Eindhoven University of Technology, The Netherlands.Google Scholar
  15. Van de Wiel, M.A., and A. Di Bucchianico (2001). Fast computation of the exact null distribution of Spearman’s rho and Page’s Lstatistic for samples with and without ties. Journal of Statistical Planning and Inference, 92, 133–145.MathSciNetCrossRefGoogle Scholar
  16. Wald, A., and J. Wolfowitz (1943). An exact test for randomness in the nonparametric case based on serial correlation. Annals of Mathematical Statistics, 14, 378–388.MathSciNetCrossRefGoogle Scholar

Copyright information

© Physica-Verlag 2001

Authors and Affiliations

  1. 1.Department of Mathematics and Computing ScienceEindhoven University of TechnologyEindhovenThe Netherlands

Personalised recommendations