c**********e 发帖数: 2007 | 1 Suppose I have a continuous variable X and a frequency variable Y. I fit a
model by SAS GENMOD. The model says that Y has a negative binomial
distribution with mean a+bX (or exp(a+bX)).
Specifically, I have a sequence X1, X2, ..., Xn and Y1, Y2, ..., Yn. How to
test Yi's follow negative binomial distributions with mean a+bXi (or exp(a+
bXi))?
Anybody has an idea? Thanks a ton. | H***a 发帖数: 735 | 2 qqplot(Y, a+bX)? Assuming you know a, b | s*********e 发帖数: 1051 | 3 周末了,可以浪费点时间。
poison and nb share the same mean function. the difference is dispersion.
therefore, the most simple is to test the significance of dispersion
parameter in nb, which should be a gamma mixture (if my memory serves me
right)
to
a+
【在 c**********e 的大作中提到】![](/moin_static193/solenoid/img/up.png) : Suppose I have a continuous variable X and a frequency variable Y. I fit a : model by SAS GENMOD. The model says that Y has a negative binomial : distribution with mean a+bX (or exp(a+bX)). : Specifically, I have a sequence X1, X2, ..., Xn and Y1, Y2, ..., Yn. How to : test Yi's follow negative binomial distributions with mean a+bXi (or exp(a+ : bXi))? : Anybody has an idea? Thanks a ton.
| s*********e 发帖数: 1051 | 4 lm test proposed by greene
http://wp.me/p42ab-cr
to
a+
【在 c**********e 的大作中提到】![](/moin_static193/solenoid/img/up.png) : Suppose I have a continuous variable X and a frequency variable Y. I fit a : model by SAS GENMOD. The model says that Y has a negative binomial : distribution with mean a+bX (or exp(a+bX)). : Specifically, I have a sequence X1, X2, ..., Xn and Y1, Y2, ..., Yn. How to : test Yi's follow negative binomial distributions with mean a+bXi (or exp(a+ : bXi))? : Anybody has an idea? Thanks a ton.
|
|