Would you help, how to initialize start value for glm estimation?
thank you.
Fo you mean initial values to help with convergence of the betas?
I don’t think that feature is available.
There is information in the documentation. I have a little experience on how to use it.
fit(GeneralizedLinearModel, X, y, d, [l = canonicallink(d)]; <keyword arguments>)
Keyword Arguments
...........
start::AbstractVector=nothing: Starting values for beta. Should have the Should have the
same length as the number of columns in the model matrix.
You are right. I’m getting an error when I try it. You should explore that on your own and file an issue if you get a persistent error.