Computational Statistics

, Volume 17, Issue 1, pp 47–58 | Cite as

The log F: A Distribution for All Seasons

  • Barry W. Brown
  • Floyd M. Spears
  • Lawrence B. Levy


Families of parametric models are widely used to summarize data, to obtain predictions, assess goodness of fit, to estimate functions of the data not easily derived directly, and to render manageable random effects. The trustworthiness of the results obtained depends on the generality of the parametric family employed. A very flexible set of statistical models based on the logarithm of an F variate was introduced over 20 years ago. It’s versatility appears to be little appreciated by the statistical community. We try to convince readers that this family belongs in the tool box of all applied statisticians and that it should be one of the first tools used in data exploration. We present examples that cover a variety of statistical functions and application areas, and we offer freely available computer code.


Generalized Log F Parametric Modeling Statistical Software Goodness of Fit Probability Distributions 


  1. Brown, B. W., Brauner, C. and Levy, L. B. (1997), Assessing Changes in the Impact of Cancer on Population Survival Without Considering Cause of Death Journal of the National Cancer Institute 89, 58–65.CrossRefGoogle Scholar
  2. Brown, B. W., Levy, L. B. (1994) Certification of Algorithm 708: Significant Digit Computation of the Incomplete Beta. ACM Transactions on Mathematical Software, 20, 393–397.CrossRefGoogle Scholar
  3. Brown, B. W., Spears, F. M., Levy, L. B., Lovato, J., and Russell, K. (1996), Algorithm 762: LLDRLF, Log-Likelihood and Some Derivatives for Log-F Models. ACM Transactions on Mathematical Software, 22, 372–382.CrossRefGoogle Scholar
  4. Di Donato, A. R., and Morris, A. H. (1992), Significant Digit Computation of the Incomplete Beta Function Ratios. ACM Transactions on Mathematical Software, 18, 360–373.CrossRefGoogle Scholar
  5. Gay, D.M. (1983), Algorithm 611. Subroutines for Unconstrained Minimization Using a Model/Trust-Region Approach. ACM Transactions on Mathematical Software 9, 503–524.MathSciNetCrossRefGoogle Scholar
  6. Hogg, S. A. and Ciampi, A. (1985), GFREG: A Computer Program for Maximum Likelihood Regression using the Generalized F Distribution. Computer Methods and Programs in Biomedicine 20, 201–215.CrossRefGoogle Scholar
  7. Kalbfleish, J. D. and Prentice, R. L. (1980),The Statistical Analysis of Failure Time Data. John Wiley and Sons, New York.Google Scholar
  8. Mack, W., Langholz, B. and Thomas, D. C. (1990), Survival Models for Familial Aggregation of Cancer. Environmental Health Perspectives 87, 27–35.CrossRefGoogle Scholar
  9. Peng, Y., Dear, K.B.G., and Denham, J.W. (1998), A Generalized F Mixture Model for Cure Rate Estimation. Statistics in Medicine 17, 813–830.CrossRefGoogle Scholar
  10. Prentice, R. L. (1975), Discrimination among Some Parametric Models. Biometrika, 62, 607–614.MathSciNetCrossRefGoogle Scholar
  11. Prentice, R. L. (1976),A Generalization of the Probit and Logit Methods for Dose Response Curves. Biometrics, 32, 761–768.CrossRefGoogle Scholar
  12. SEER (1997), Surveillance, Epidemiology, and End Results public use CDROM (1973–1995). National Cancer Institute, Division of Cancer Prevention and Control, Surveillance Program, Cancer Statistics Branch.Google Scholar
  13. Taylor, B. (1998), Stock Market Indices, 1800–1995. URL Used with permission.
  14. U.S.A. Today (1998), AL team-by-team statistics. URL Used with permission.

Copyright information

© Physica-Verlag 2002

Authors and Affiliations

  • Barry W. Brown
    • 1
  • Floyd M. Spears
    • 2
  • Lawrence B. Levy
    • 1
  1. 1.Department of BiomathematicsUniversity of Texas M. D. Anderson Cancer CenterHoustonUSA
  2. 2.University of Houston-Clear LakeHoustonUSA

Personalised recommendations