Families

Each family is named for the parameters it models, and each parameter has a helper method on the family (location, scale, mu, phi, and so on) that takes the declarations on the Declarations page and returns the predictor for that parameter.

• GaussianLS: location and scale of a Gaussian response.

• GammaLS: mean and coefficient of variation of a positive response.

• LogNormalLS: mean (or location) and scale of a log-normal response.

• NegativeBinomialLS: mean and NB2 size of a count response.

• GeneralizedGammaLSS: mean (or location), scale and shape of a generalized gamma, which nests the log-normal, Weibull and gamma.

• GeneralizedParetoLSS: scale and shape of threshold excesses.

• TweedieLSS: mean, dispersion and variance power of a nonnegative response with a point mass at zero.

• TwoPieceLogNormalLSS: mean (or location), scale and skew of a two-piece log-normal.

• TwoPieceNormalLSS: location, scale and skew of a two-piece normal on the real line.

superglm.GaussianLS	Gaussian responses with separate mean and standard-deviation predictors.
superglm.GammaLS	Positive responses with mean and coefficient-of-variation predictors.
superglm.LogNormalLS	Log-normal with natural parameters (mean | location, scale).
superglm.NegativeBinomialLS	Negative-binomial responses with mean and size predictors.
superglm.GeneralizedGammaLSS	Generalized gamma with natural parameters (mean | location, scale, shape).
superglm.GeneralizedParetoLSS	Threshold excesses with generalized Pareto scale and shape predictors.
superglm.TweedieLSS	Nonnegative responses with mean, dispersion and variance-power predictors.
superglm.TwoPieceLogNormalLSS	Two-piece log-normal with natural parameters (mean | location, scale, skew).
superglm.TwoPieceNormalLSS	Epsilon-skew two-piece normal on the real line, (location, scale, skew).