Some Approximations for the Noncentral-F Distribution
- 1 August 1976
- journal article
- research article
- Published by JSTOR in Technometrics
- Vol. 18 (3) , 351
- https://doi.org/10.2307/1268745
Abstract
Among the various approximations for the noncentral-F distribution. the best known are the computationally simpler approximation due to Severo and Zelen [18] and Laubscher [9], which requires only the normal distribution, and a more accurate approximation due to Tiku [21], which requires the F distribution. Thcsr approximations are systematically evaluated. Some variations of these approximations are considered and found to he of comparable accuracy. An Edgeworth-acries approximation, which requires only the normal distribution, is developed and sren to be superior in accuracy. A FORTRAN-subroutine for the Edgeworth-series approximation is given.Keywords
This publication has 0 references indexed in Scilit: