I am trying to estimate the value of the Tweedie p-index parameter using the tweedie.profile function in the tweedie package in R. An example of the function being used is given below
model.out <- tweedie.profile( formula = y1~ x1+ x2+ x3+ x4-1,
p.vec = seq(1.2,1.9, by = 0.1), link.power = 0, data = train.data,
weights = w1^p-1, offset = log(w1), fit.glm = TRUE,
do.smooth = TRUE, do.plot = TRUE, do.ci = TRUE, conf.level = 0.95)
The the model is designed to have the weight be the w1 ^ (p-index parameter (p) - 1) however since I don't know the p index parameter I would like to set p to be the current value p value that tweedie.profile is using to compute the maximum likelihood, it it possible to call this value e.g. as w1^p.vec$value-1? Thanks for any help
