OVERDISPERSION IN CLINICAL MASTITIS DATA FROM DAIRY HERDS - A NEGATIVE BINOMIAL APPROACH

被引:29
作者
SCHUKKEN, YH
CASELLA, G
VANDENBROEK, J
机构
[1] STATE UNIV UTRECHT,DEPT HERD HLTH & REPROD,3584 CL UTRECHT,NETHERLANDS
[2] CORNELL UNIV,NEW YORK STATE COLL AGR & LIFE SCI,DEPT PLANT BREEDING & BIOMETRY,BIOMETR UNIT,ITHACA,NY 14853
[3] STATE UNIV UTRECHT,CTR BIOSTAT,3584 CL UTRECHT,NETHERLANDS
关键词
D O I
10.1016/0167-5877(91)90007-O
中图分类号
S85 [动物医学(兽医学)];
学科分类号
0906 ;
摘要
Cases of clinical mastitis on dairy farms were modelled using a negative binomial distribution. The negative binomial distribution is a gamma mixture of Poisson distributions, hence it is possible that the number of cases per farm had a Poisson distribution and the mean of these Poisson distributions had a gamma distribution. Risk factors associated with the number of cases per farm can be modelled with a negative binomial distribution but calculations involving the negative binomial likelihood are not straightforward. However, a very close approximation to the negative binomial is the Poisson model with an extra parameter for overdispersion. Fitting the Poisson model with overdispersion is illustrated.
引用
收藏
页码:239 / 245
页数:7
相关论文
共 14 条