Extended techniques for generalized linear models (GLMs), especially for binary responses, including parametric links and heteroskedastic latent variables.
