Extended moment series and the parameters of the negative binomial distribution

Journal Article · Thu Mar 01 00:00:00 EST 1984 · Biometrics; (United States)

DOI:https://doi.org/10.2307/2530767· OSTI ID:6749174

Bowman, K O

Recent studies indicate that, for finite sample sizes, moment estimators may be superior to maximum likelihood estimators in some regions of parameter space. In this paper a statistic based on the central moment of the sample is expanded in a Taylor series using 24 derivatives and many more terms than previous expansions. A summary algorithm is required to find meaningful approximants using the higher-order coefficients. A example is presented and a comparison between theoretical assessment and simulation results is made.

• Research Organization: Union Carbide Corp., Oak Ridge, TN
• DOE Contract Number: W-7405-ENG-26
• OSTI ID: 6749174
• Journal Information: Biometrics; (United States), Vol. 40
• Country of Publication: United States
• Language: English

Related Subjects

71 CLASSICAL AND QUANTUM MECHANICS
GENERAL PHYSICS
STATISTICS
DISTRIBUTION FUNCTIONS
MAXIMUM-LIKELIHOOD FIT
SERIES EXPANSION
FUNCTIONS
MATHEMATICS
NUMERICAL SOLUTION
657000* - Theoretical & Mathematical Physics