I'm trying to understand GLM, and was wondering why one should want to use an identity link function (for example with Poisson regression). With the identity link, isn't the GLM basically the same as linear regression? The only difference, as far as I understand it, is that the regression parameters are estimated by maximum likelihood in GLM and with least squares in linear regression. Is that right? Or is there something I'm missing?
Any help is appreciated,
Advertise on Talk Stats