Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Moment series for the coefficient of variation in Weibull sampling

Conference ·
OSTI ID:5229608
For the 2-parameter Weibull distribution function F(t) = 1 - exp(-t/b)/sup c/, t > 0, with c and b positive, a moment estimator c* for c is the solution of the equationGAMMA(1 + 2/c*)/GAMMA/sup 2/ (1 + 1/c*) = 1 + v*/sup 2/ where v* is the coefficient of variation in the form ..sqrt..m/sub 2//m/sub 1/', m/sub 1/' being the sample mean, m/sub 2/ the sample second central moment (it is trivial in the present context to replace m/sub 2/ by the variance). One approach to the moments of c* (Bowman and Shenton, 1981) is to set-up moment series for the scale-free v*. The series are apparently divergent and summation algorithms are essential; we consider methods due to Levin (1973) and one, introduced ourselves (Bowman and Shenton, 1976).
Research Organization:
Oak Ridge National Lab., TN (USA); Georgia Univ., Athens (USA). Office of Computing and Information Service
DOE Contract Number:
W-7405-ENG-26
OSTI ID:
5229608
Report Number(s):
CONF-810842-11-Draft; ON: DE84007591
Country of Publication:
United States
Language:
English