一个类似的问题被问及β二项分布的alpha beta估计。
我理解那里的答案;我的问题更多的是关于我得到的错误。
我的数据是:
r1 = c(3, 1, 2, 6, 3, 6, 0, 4,2, 4, 6)
r2 = c(5, 1, 0, 1, 5, 5, 2, 2, 2, 1, 7)
z = cbind(r1,r2)
fit = vglm(z ~ 1, betabinomialff, trace = TRUE)
我收到以下错误消息:
VGLM linear loop 1 : loglikelihood = -17.569074
VGLM linear loop 2 : loglikelihood = -16.796047
VGLM linear loop 3 : loglikelihood = -16.556079
VGLM linear loop 4 : loglikelihood = -16.489902
VGLM linear loop 5 : loglikelihood = -16.486634
VGLM linear loop 6 : loglikelihood = -16.486634
Taking a modified step....................
Warning messages:
1: In checkwz(wz, M = M, trace = trace, wzepsilon = control$wzepsilon) :
22 diagonal elements of the working weights variable 'wz' have been replaced by 1.819e-12
2: In vglm.fitter(x = x, y = y, w = w, offset = offset, Xm2 = Xm2, :
迭代因为半步长非常小而终止 3:在vglm.fitter中(x = x,y = y,w = w,offset = offset,Xm2 = Xm2,: 由于半步收敛,一些量如z,残差,SE可能不准确
我如何获得beta二项式模型的alpha beta估算值,而不会遇到此错误?我对使用vglm函数的所有细节都不是很熟悉,所以感谢任何帮助。
谢谢!