Logo

statsmodels.genmod.families.family.Gamma

class statsmodels.genmod.families.family.Gamma(link=<class 'statsmodels.genmod.families.links.inverse_power'>)[source]

Gamma exponential family distribution.

Parameters:

link : a link instance, optional

The default link for the Gamma family is the inverse link. Available links are log, identity, and inverse. See statsmodels.family.links for more information.

Attributes

Gamma.link a link instance The link function of the Gamma instance
Gamma.variance varfunc instance variance is an instance of statsmodels.family.varfuncs.mu_squared

Methods

deviance(endog, mu[, scale]) Gamma deviance function
fitted(lin_pred) Fitted values based on linear predictors lin_pred.
loglike(endog, mu[, scale]) Loglikelihood function for Gamma exponential family distribution.
predict(mu) Linear predictors based on given mu values.
resid_anscombe(endog, mu) The Anscombe residuals for Gamma exponential family distribution
resid_dev(endog, mu[, scale]) Gamma deviance residuals
starting_mu(y) Starting value for mu in the IRLS algorithm.
weights(mu) Weights for IRLS steps

Previous topic

statsmodels.genmod.families.family.Binomial.weights

Next topic

statsmodels.genmod.families.family.Gamma.deviance

This Page